Hewlett Packard Enterprise 10.00 Data Protector Integration Guide
Data Protector 10.00 is a software that enables zero downtime backup integration. It allows you to perform backups without interrupting the operation of the application system. This is achieved by creating replicas of the data that is being backed up.
Advertisement
Advertisement
Data Protector
Software Version: 10.00
Zero Downtime Backup Integration Guide
Document Release Date: June 2017
Software Release Date: June 2017
Zero Downtime Backup Integration Guide
Legal Notices
Warranty
The only warranties for Hewlett Packard Enterprise Development LP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HPE shall not be liable for technical or editorial errors or omissions contained herein.
The information contained herein is subject to change without notice.
Restricted Rights Legend
Confidential computer software. Valid license from HPE required for possession, use or copying. Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation, and
Technical Data for Commercial Items are licensed to the U.S. Government under vendor's standard commercial license.
Copyright Notice
© Copyright 2017 Hewlett Packard Enterprise Development LP
Trademark Notices
Adobe™ is a trademark of Adobe Systems Incorporated.
Microsoft® and Windows® are U.S. registered trademarks of Microsoft Corporation.
UNIX® is a registered trademark of The Open Group.
This product includes an interface of the 'zlib' general purpose compression library, which is Copyright ©
1995-2002 Jean-loup Gailly and Mark Adler.
Documentation updates
The title page of this document contains the following identifying information: l
Software Version number, which indicates the software version.
l
Document Release Date, which changes each time the document is updated.
l
Software Release Date, which indicates the release date of this version of the software.
To check for recent software updates, go to https://softwaresupport.hpe.com/patches .
To verify that you are using the most recent edition of a document, go to https://softwaresupport.hpe.com/manuals .
This site requires that you register for an HPE Passport and sign in. To register for an HPE Passport ID, go to https://hpp12.passport.hpe.com/hppcf/login.do
.
You will also receive updated or new editions if you subscribe to the appropriate product support service.
Contact your HPE sales representative for details.
Support
Visit the HPE Software Support Online web site at https://softwaresupport.hpe.com
.
This web site provides contact information and details about the products, services, and support that HPE
Software offers.
HPE Software online support provides customer self-solve capabilities. It provides a fast and efficient way to access interactive technical support tools needed to manage your business. As a valued support customer, you can benefit by using the support web site to: l
Search for knowledge documents of interest l
Submit and track support cases and enhancement requests
Data Protector (10.00) Page 2 of 413
Zero Downtime Backup Integration Guide l
Download software patches l
Access product documentation l
Manage support contracts l
Look up HPE support contacts l
Review information about available services l
Enter into discussions with other software customers l
Research and register for software training
Most of the support areas require that you register as an HPE Passport user and sign in. Many also require a support contract.
To register for an HPE Passport ID, go to https://hpp12.passport.hpe.com/hppcf/login.do
.
To find more information about access levels, go to https://softwaresupport.hpe.com/web/softwaresupport/access-levels .
Data Protector (10.00) Page 3 of 413
Zero Downtime Backup Integration Guide
Contents
Chapter 1: Data Protector Oracle Server ZDB integration
Oracle backup set ZDB concepts
Oracle proxy-copy ZDB concepts
Configuring the HPE P6000 EVA Disk Array Family in ASM environments
Configuring the HPE P9000 XP Disk Array Family in ASM environments
Linking Oracle Server with the Data Protector MML
Oracle 12c CDB and PDB mode support
Configuring Oracle user accounts
Configuring Oracle operating system user accounts
Configuring Oracle database user accounts
Checking configuration for instant recovery
Switching between Oracle backup methods
Creating backup specifications
Examples of pre-exec and post-exec scripts on UNIX systems
Editing the Oracle RMAN script
Data Protector (10.00) Page 4 of 413
15
Zero Downtime Backup Integration Guide
Starting a backup using the GUI
Starting a backup using the CLI
Restoring from backup media to the application system on LAN
Restoring Oracle using the Data Protector GUI
Restoring database items in a disaster recovery
Restoring the recovery catalog database
Restoring Oracle database objects
Restoring tablespaces and datafiles
Duplicating an Oracle database
Restore, recovery, and duplicate options
Preparing the Oracle database for restore
Connection strings used in the examples
Example of full database restore and recovery
Example of point-in-time restore
Example of tablespace restore and recovery
Example of datafile restore and recovery
Example of archive log restore
Restoring using another device
Restoring Oracle to different cell/client
Instant recovery and database recovery
Instant recovery using the Data Protector GUI
Oracle database recovery after the instant recovery
Oracle in Veritas Cluster instant recovery
Chapter 2: Data Protector SAP R/3 ZDB integration 124
Data Protector (10.00) Page 5 of 413
Zero Downtime Backup Integration Guide
Data Protector SAP R/3 configuration file
Configuring SQL*Net V2 or Net8 TNS listener
Sharing directories on the application system
Configuring the SAP R/3 parameter file
Creating backup specifications
Modifying backup specifications
What happens during the preview?
Configuring SAP compliant ZDB sessions
Data Protector (10.00) Page 6 of 413
Zero Downtime Backup Integration Guide
Instant recovery using the Data Protector GUI
Instant recovery using the Data Protector CLI
Instant recovery from replicas containing the control file
Restoring using another device
Using the Data Protector CLI or SAP commands
Prerequisites on the SAP side of the integration
Prerequisites on the SAP side of the integration
Verifying the prerequisites (Oracle side)
Verifying the prerequisites (SAP side)
Verifying the backup configuration
Configuration and backup problems
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Data Protector SQL Server configuration file
Configuring an SQL Server cluster
Configuring SQL Server instances
Data Protector (10.00) Page 7 of 413
193
Zero Downtime Backup Integration Guide
Changing and checking configuration
SQL Server-specific backup options
Restoring to a different SQL Server instance or/and different SQL Server
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Windows domain user account for backup and restore sessions
User account for executing Exchange Management cmdlet operations
Microsoft Exchange Server backup types
Data Protector (10.00) Page 8 of 413
228
Zero Downtime Backup Integration Guide
Replica rotation in DAG environments
Object operations considerations
Creating backup specifications
Modifying backup specifications
What happens during the preview?
Repair all passive copies with failed status
Restore to a new mailbox database
Restore files to a temporary location
Restoring to a standalone database
Restoring data to a new database
Restoring data to a temporary location
Instant recovery in DAG environments
Restore chain during instant recovery
Finding information for restore
Restoring using the Data Protector GUI
Restoring using the Data Protector CLI
Restoring using another device
Performing instant recovery using the Data Protector GUI
Performing instant recovery using the Data Protector CLI
Data Protector (10.00)
Page 9 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Installation and configuration
Microsoft Office SharePoint Server 2007
Microsoft SharePoint Server 2010
Microsoft SharePoint Server 2010
Microsoft SharePoint Server 2013
Creating backup specifications (examples)
Modifying backup specifications
Starting backup sessions (examples)
Restoring using the Data Protector GUI
Restoring using the Data Protector CLI
285
Data Protector (10.00) Page 10 of 413
Zero Downtime Backup Integration Guide
After restore, you cannot connect to the Central Administration webpage 305
Backup fails with the error Failed to resume Service Windows SharePoint Services Help
After restore, a quiesce operation fails 306
307 After restore, you cannot connect to the FAST Search Server
The SharePoint_VSS_backup.ps1 script stops responding and the farm stays in read only mode
SharePoint Search service application not operational after restore
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware 310
Virtual machine templates vStorage Image backup method
Non-Changed Block Tracking (Non-CBT) backup
Considerations for Quiescence Operations
Restore of VMware objects backed up with vStorage Image method
Restore of Nova Instances and Shadow VMs backed up with vStorage Image +
Data Protector (10.00) Page 11 of 413
Zero Downtime Backup Integration Guide
StoreOnce Recovery Manager Central Integration
Importing and configuring VMware clients
Changing the configuration of VMware clients
Checking the configuration of VMware clients
Customizing the Data Protector behavior with omnirc options
Adding the RMC Server details in Data Protector using the Command Line Interface
vStorage Image + OpenStack backup method limitations
Creating backup specifications
Creating backup specifications for RMC backups
Modifying backup specifications
What happens during the preview?
Preparing for disaster recovery
vStorage Image + OpenStack restore limitations
Power On and Live Migrate limitations
Finding information for restore
Restoring using the Data Protector GUI
Restoring using the Data Protector CLI
Recovering virtual machines manually
Recovering virtual machines after restore to a directory
Recovering with the VM configuration file in the VMX format
Data Protector (10.00) Page 12 of 413
Zero Downtime Backup Integration Guide
Recovering with the VM configuration file in the XML format
Recovering virtual machines after restore to a datacenter
Restoring using another device
Cleaning up a datastore after a failed restore
Chapter A: Appendix
Reconfiguring an Oracle instance for instant recovery
Examples for moving the control files and redo logs to different locations
ZDB integrations omnirc options
Send documentation feedback
405
413
Data Protector (10.00) Page 13 of 413
Zero Downtime Backup Integration Guide
Data Protector (10.00) Page 14 of 413
Chapter 1: Data Protector Oracle Server ZDB integration
Introduction
You can employ a variety of backup strategies to best meet your system priorities. If database availability is the highest priority, for instance, your backup strategy should include online backups that are performed frequently to minimize recovery time. This strategy limits downtime, but uses system resources more intensively. The Data Protector zero downtime backup (ZDB) functionality offers online backup capabilities with minimal degradation of the application system performance.
Supported disk arrays
The following disk arrays can be used for zero downtime backup (ZDB) of the Oracle Server data: l
HPE P6000 EVA Disk Array Family (P6000 EVA Array) l
HPE P9000 XP Disk Array Family (P9000 XP Array) l
EMC Symmetrix (EMC) l
Non–HPE Storage Arrays (NetApp storage, EMC VNX, and EMC VMAX storage families)
NOTE:
With the Data Protector EMC and Data Protector non–HPE Storage Array integrations, instant recovery is not supported, and ZDB to tape is the only supported ZDB form.
In Oracle Server configurations that use Automatic Storage Management (ASM), zero downtime backup and instant recovery are supported with the Data Protector P6000 EVA Array and Data
Protector P9000 XP Array integrations.
Advantages
The advantages of using Data Protector Oracle ZDB integration are: l
ZDB reduces the performance degradation of the application system.
l
The tablespaces are in backup mode (online backup) or the database is shut down (offline backup) only during the short period required to create a replica (split the mirror disks or create snapshots).
l
The load to the application system is significantly reduced. Following the replica creation, tape backup can be started on the copied data, at leisure, using a separate backup system.
The Data Protector Oracle ZDB integration offers online and offline backup of your Oracle Server System
(application system).
The online backup concept is widely used since it enables high application availability. Offline backup requires shutting down the database while creating a replica, and therefore does not offer high availability.
ZDB methods and Oracle versions
The installation, upgrade, configuration, and parts of backup flow are different depending on the selected
Oracle ZDB method. These differences are indicated where appropriate.
Data Protector (10.00) Page 15 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
The procedures for configuration of backup specifications and starting or scheduling backups are the same, regardless of the Oracle ZDB method.
Backup and restore types
Backup l
Online ZDB to disk, ZDB to tape, and ZDB to disk+tape.
During the creation of a replica, the database on the application system is in hot backup mode. If a
ZDB-to-tape or a ZDB-to-disk+tape session is being performed, the streaming of the data to tape media is subsequently performed on the backup system.
l
Offline ZDB to disk, ZDB to tape, and ZDB to disk+tape.
During the creation of a replica, the database is shut down on the application system. Therefore, the database is not available during the short time that it takes to create the replica. If a ZDB-to-tape or a
ZDB-to-disk+tape session is being performed, the streaming of the data to tape media is subsequently performed on the backup system.
With both online and offline ZDB to tape or ZDB to disk+tape, a standard Data Protector (non-ZDB) backup of the recovery catalog and the control file is started automatically, after the target database backup is finished on the backup system. However, you can disable this when creating a backup specification.
NOTE:
Backup of the recovery catalog and control file is not performed with ZDB to disk.
The Oracle Recovery Manager utility (RMAN) is not aware of ZDB-to-disk sessions.
IMPORTANT:
Backup of archived logs cannot be done with the Data Protector Oracle ZDB integration.
Backup of archive logs and control file has to be done following the standard Data Protector
Oracle integration backup procedure. For more information on Oracle archive log backup with
Data Protector see the HPE Data Protector Integration Guide.
NOTE:
On EMC, decision support, application testing, and similar tasks are possible only if the Oracle binaries are installed on the backup system as well. In most cases, however, the Data
Protector EMC integration requirement is that application binaries are installed on the application system only.
Restore
Using Data Protector and the disk array integrations, you can perform the following types of restore: l
Restoring from backup media to the application system on LAN (standard Data Protector restore) and using RMAN on the application system, you can: o recover a whole database o recover a part of a database o recover a whole database as it was at a specific point in time
Data Protector (10.00) Page 16 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Using the instant recovery functionality and RMAN on the application system, you can: o perform a full database restore and database recovery o perform recovery from incremental backup (for ZDB to tape or ZDB to disk+tape) o perform recovery from a chain of incremental backups (for ZDB to tape or ZDB to disk+tape) o restore a datafile to a location other than its original one
Oracle recovery methods, below provides an overview of recovery methods, depending on the type of
backup that was performed and type of recovery required.
Oracle recovery methods
Disk array Backup types Recover the whole database until Recover a part of database
Now A point in time, until now logseq/thread, or SCN number
Restore Restore Restore
P9000 XP,
P6000
EVA,
EMC, non–HPE
Storage
Arrays
ZDB to tape - online
ZDB to tape - offline
P9000 XP,
P6000
EVA
ZDB to disk - online
Restore
Instant recovery+ database recovery
Restore
Instant recovery+ database recovery
1
Restore
N/A
ZDB to disk - offline
ZDB to disk+tape - online
Instant recovery Instant recovery+ database recovery
2 l
Restore or l
Instant recovery+ database l
Restore or l
Instant recovery+ database
N/A
Restore
1
The database must be put in archive mode
2
The database must be put in archive mode
Data Protector (10.00) Page 17 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Disk array Backup types
ZDB to disk+tape - offline
Recover the whole database until Recover a part of database
Now A point in time, until now logseq/thread, or SCN number
recovery recovery l
Restore or l
Instant recovery l
Restore or l
Instant recovery+ database recovery
1
Restore
Legend
Restore
Instant recovery + database recovery
Use the Data Protector GUI or RMAN scripts to restore the database from backup media to the application system on LAN.
The following three options are possible: l
Perform instant recovery followed by database recovery from the Data
Protector Instant Recovery GUI context or l
Perform instant recovery first and then perform database recovery from the Data Protector Restore GUI context or l
Perform instant recovery first and then use RMAN scripts to recover the database.
Perform instant recovery without database recovery.
Instant recovery
See the HPE Data Protector Concepts Guide for an overview of ZDB concepts and terminology.
Integration concepts
The Data Protector Oracle integration links the Oracle database management software with Data
Protector. From the Oracle point of view, Data Protector represents a media management software. On the other hand, the Oracle database management system can be seen as a data source for backup, using media controlled by Data Protector.
Components
The software components involved in backup and restore processes are:
1
The database must be put in archive mode
Data Protector (10.00) Page 18 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
The Oracle Recovery Manager (RMAN) l
The Data Protector Oracle integration software
Integration functionality overview
The Data Protector Oracle Integration agent ( ob2rman.pl
) works with RMAN to manage all aspects of the following operations on the Oracle target database: l
Database startup and shutdown l
Backups (backup and copy) l
Recovery (restore, recovery, and duplication)
How does the integration work?
ob2rman.pl
executes RMAN, which directs the Oracle server processes on the target database to perform backup, restore and recovery. RMAN maintains the required information about the target databases in the recovery catalog, the Oracle central repository of information, and in the control file of a particular target database.
The main information which ob2rman.pl
provides to RMAN is: l
Number of allocated RMAN channels l
RMAN channel environment parameters l
Information on the database objects to be backed up or restored
For backup, ob2rman.pl
uses the Oracle target database views to get information on which logical
(tablespaces) and physical (datafiles) target database objects are available for backup.
For restore, ob2rman.pl
uses current control file or recovery catalog (if used) to get information on which objects are available for restore.
Using the Data Protector integration with RMAN, you can back up and restore the Oracle control files, datafiles, and archived redo logs.
The interface from the Oracle server processes to Data Protector is provided by the Data Protector
Oracle integration Media Management Library (MML), which is a set of routines that allows the reading and writing of data to General Media Agents.
Besides handling direct interaction with the media devices, Data Protector provides scheduling, media management, network backups, monitoring, and interactive backup.
A backup that includes all datafiles and current control file that belong to an Oracle Server instance is known as a whole database backup.
These features can be used for online or offline backup of the Oracle target database. However, you must ensure that the backup objects (such as tablespaces) are switched into the appropriate state before and after a backup session. For online backup, the database instance must operate in the
ARCHIVELOG mode; whereas for offline backup, objects need to be prepared for backup using the Preexec and Post-exec options in the backup specification.
The Data Protector backup specification contains information about backup options, commands for
RMAN, Pre- and Post-exec commands, media, and devices.
The Data Protector backup specification allows you to configure a backup and then use the same specification several times. Furthermore, scheduled backups can only be performed using a backup specification.
Data Protector (10.00) Page 19 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Backup and restore of an Oracle target database can be performed using the Data Protector User
Interface or the RMAN utility.
The heart of the Data Protector Oracle integration is MML, which enables an Oracle server process to issue commands to Data Protector for backing up or restoring parts or all of the Oracle target database files. The main purpose is to control direct interaction with media and devices.
Non-ZDB flow
A Data Protector scheduled or interactive backup is triggered by the Data Protector Backup Session
Manager, which reads the backup specification and starts the ob2rman.pl
command on the Oracle
Server under the operating system user account specified in the backup specification. Further on, ob2rman.pl
prepares the environment to start the backup, and issues the RMAN backup command.
RMAN instructs the Oracle Server processes to perform the specified command.
The Oracle Server processes initialize the backup through MML, which establishes a connection to the
Data Protector Backup Session Manager. The Backup Session Manager starts the General Media
Agent, sets up a connection between MML and the General Media Agent, and then monitors the backup process.
The Oracle Server processes read the data from the disks and send it to the backup devices through
MML and the General Media Agent.
RMAN writes information regarding the backup either to the recovery catalog (if one is used) or to the control file of the Oracle target database.
Messages from the backup session are sent to the Backup Session Manager, which writes messages and information regarding the backup session to the IDB.
The Data Protector General Media Agent writes data to the backup devices.
Restore flow
A restore session can be started using: l
Data Protector GUI l
RMAN CLI
You must specify which objects are to be restored.
A restore from the Data Protector user interface is triggered by the Data Protector Restore Session
Manager, which starts the ob2rman.pl
command. ob2rman.pl
prepares the environment to start the restore, and issues the RMAN restore command. RMAN checks the recovery catalog (if one is used) or the control file to gather the information about the Oracle backup objects. It also contacts the Oracle
Server processes, which initialize the restore through MML. MML establishes a connection with the
Restore Session Manager and passes along the information about which objects and object versions are needed.
The Restore Session Manager checks the IDB to find the appropriate devices and media, starts the
General Media Agent, establishes a connection between MML and the General Media Agent, and then monitors the restore and writes messages and information regarding the restore to the IDB.
The General Media Agent reads the data from the backup devices and sends it to the Oracle Server processes through MML. The Oracle Server Processes write the data to the disks.
Data Protector (10.00) Page 20 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Data Protector Oracle integration concept
Database files can also be managed by Automatic Storage Management (ASM).
Legend
SM
The Data Protector Session Manager, which can be the Data Protector Backup Session
Manager during a backup session and the Data Protector Restore Session Manager during a restore session.
RMAN
The Oracle Recovery Manager.
Data
Protector
MML
The Data Protector Oracle integration Media Management Library, which is a set of routines that enables data transfer between the Oracle Server and Data Protector.
The Oracle-defined application programming interface.
Backup
API
IDB
MA
The Data Protector Internal Database where all the information about Data Protector sessions, including session messages, objects, data, and used devices and media, is written.
The Data Protector General Media Agent, which reads and writes data from and to media devices.
Data Protector (10.00) Page 21 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Oracle backup set ZDB concepts
See the HPE Data Protector Concepts Guide for a general description of ZDB-to-disk, ZDB-to-tape, and ZDB-to-disk+tape and instant recovery concepts.
With the Oracle backup set ZDB method, the entire data to be backed up is provided to Data Protector through the Oracle API—the data is streamed through the Data Protector Oracle integration MML.
Depending on the location of the Oracle control file, online redo log files, and SPFILE, the following two options are possible: l
Oracle control file, online redo log files, and SPFILE reside on a different volume group (if LVM is used) or source volume than Oracle datafiles.
By default, instant recovery for such a configuration is enabled.
l
Oracle control file, online redo log files, SPFILE reside on the same volume group (if LVM is used) or source volume as Oracle datafiles.
By default, instant recovery for such a configuration is not enabled. You can enable instant recovery by setting the ZDB_ORA_INCLUDE_CF_OLF , ZDB_ORA_INCLUDE_SPF , and ZDB_ORA_NO_CHECKCONF_
IRomnirc options to 1
. See ZDB integrations omnirc options, on page 408 .
IMPORTANT:
If you enable instant recovery by setting the above mentioned options, note that the control file, SPFILE, and online redo logs are overwritten during instant recovery.
The Oracle archived redo log files do not have to reside on source volumes.
Oracle backup set ZDB concept
Data Protector (10.00) Page 22 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Oracle backup set ZDB concept, on the previous page presents only the default integration behavior,
where Oracle control file, online redo log files, and SPFILE reside on a different volume group (if LVM is used) or source volume than Oracle datafiles. Oracle database files can also be managed by ASM,
.
.
Legend
MA
SM
The General Media Agent writes data from a replica to backup media. The General Media
Agent typically resides on the backup system.
The session manager controls backup and restore sessions and writes session information to the IDB.
Disk
Array
Agent
The disk array agents (ZDB agents) are SYMA (for EMC), SSEA (for P9000 XP Array), and SMISA (for P6000 EVA Array and non–HPE Storage Arrays).
Data
Protector
MML
The Data Protector Oracle integration Media Management Library, which is a set of routines that enables data transfer between the Oracle Server and Data Protector. This is a Data Protector software library that is linked to the Oracle software.
Data Protector (10.00) Page 23 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Backup process
Oracle backup set ZDB flow
NOTE:
ZDB agents are SYMA for EMC, SSEA for P9000 XP Array, and SMISA for P6000 EVA Array and non–HPE Storage Arrays.
See the HPE Data Protector Concepts Guide for a general description of ZDB and instant recovery concepts.
Data Protector (10.00) Page 24 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
See the HPE Data Protector Zero Downtime Backup Administrator's Guide for a general description of the ZDB-to-disk, ZDB-to-tape, and ZDB-to-disk+tape session flows and for the explanation of actions triggered by ZDB options.
This section provides only the information relevant to the Data Protector Oracle ZDB integration.
Operations on a replica (mounting, activating volume/disk groups,...) described below are dependent on or triggered by ZDB options. See the HPE Data Protector Zero Downtime Backup Administrator's
Guide for more information on these options.
l
Data Protector executes the ob2rman.pl
command on the backup system. This command retrieves a list of files or disk images to be backed up from the Oracle database on the application system and starts the resolving process. The list is used only to determine the source volumes to be replicated.
If the location for control file copy is specified during configuration, ob2rman.pl
makes a copy of the control file to the specified directory on the application system. This directory has to reside on a disk array source volume.
l
When performing an online ZDB session, ob2rman.pl
then sets the Oracle target database into backup mode by issuing the sqlplus command “ALTER TABLESPACE BEGIN BACKUP ”, starts the procedure to create a replica of the source volumes on which the database is installed; and after the replica is created, takes the database out of backup mode by issuing the sqlplus command “ALTER
TABLESPACE END BACKUP ”.
When performing an offline ZDB session, ob2rman.pl
shuts down the Oracle database, starts the procedure to create a replica of the source volumes on which the database is installed; and after the replica is created, starts up the Oracle database.
l ob2rman.pl
then starts the procedure to prepare the replica on the backup system. In this step, volume/disk groups on the backup system are enabled and, unless the database is installed on raw partitions, the mount points with the Oracle database files are mounted.
l
A ZDB agent then mounts the database on the backup system to the mount points with the same names (created by Data Protector) as on the application system.
NOTE:
There must be nothing already mounted on the mount point concerned on the backup system, or the resolving and backup will fail.
l
If a ZDB-to-disk session is being performed, at this point the remaining ZDB options are processed and details of the session are written to the ZDB database. The session then finishes. The following steps in this description are not performed, therefore RMAN is not given any information about ZDBto-disk session.
l
If a ZDB-to-tape or a ZDB-to-disk+tape session is being performed, the processing continues as follows: o ob2rman.pl
starts the Oracle backup command RMAN on the backup system, and then sends the Oracle RMAN Backup Command Script to the RMAN cmdfile (input command file).
o
RMAN contacts the Oracle database instance on the backup system, which contacts Data
Protector via SBT API and initiates a backup.
o
The Oracle database instance on the backup system reads data from the replica and sends it to the Data Protector General Media Agent for writing to the backup device.
o
At the end of data transfer, the backup system is disabled (filesystems are dismounted on all
Data Protector (10.00) Page 25 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration platforms and volume/disk groups deactivated on UNIX systems) and links are re-established.
o
The recovery catalog and the control file are backed up automatically after the target database backup is finished on the backup system. However, you can disable this when creating a backup specification.
NOTE:
A replica of the archive logs is not created; therefore, the archive logs should be backed up from the application system, following the standard Data Protector Oracle archive logs backup procedure.
Oracle proxy-copy ZDB concepts
See the HPE Data Protector Concepts Guide for a general description of ZDB-to-disk, ZDB-to-tape,
ZDB-to-disk+tape, and instant recovery concepts.
The Data Protector Oracle integration MML supports the Proxy Copy functionality. This enables Data
Protector to perform backup using filesystem backup methods.
Depending on the location of the Oracle control file, online redo log files, and SPFILE, the following two options are possible: l
Oracle control file, online redo log files, and SPFILE reside on a different volume group (if LVM is used) or source volume than Oracle datafiles.
By default, instant recovery is enabled if this option is selected in the GUI.
l
Oracle control file, online redo log files, and SPFILE reside on the same volume group (if LVM is used) or source volume as Oracle datafiles.
By default, instant recovery is not enabled, even if this option is selected in the GUI. You can enable instant recovery by setting the ZDB_ORA_INCLUDE_CF_OLF , ZDB_ORA_INCLUDE_SPF , and ZDB_ORA_
NO_CHECKCONF_IRomnirc options to 1
. See ZDB integrations omnirc options, on page 408
.
IMPORTANT:
If you enable instant recovery by setting the above mentioned options, note that the control file, SPFILE, and online redo logs are overwritten during instant recovery.
The Oracle archived redo log files do not have to reside on source volumes.
Oracle proxy-copy ZDB concept, below shows the architecture of the Data Protector Oracle ZDB
integration. The figure illustrates the configuration, in which the backup is performed on the backup system. It presents the default integration behavior, where Oracle control file, online redo log files, and
SPFILE reside on different disk array source volumes than the Oracle data files. For more information
Oracle proxy-copy ZDB concept
Data Protector (10.00) Page 26 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
MA
SM
Disk Array
Agent
MML
The General Media Agent writes data from a replica to backup media. The General Media Agent typically resides on the backup system.
The session manager controls backup and restore sessions and writes the session information to the IDB.
The disk array agents (ZDB agents) are SYMA (for EMC),
SSEA (for P9000 XP Array), and SMISA (for P6000 EVA Array and non–HPE Storage Arrays).
The Data Protector Oracle integration Media Management
Library, which is a set of routines that enables data transfer between the Oracle Server and Data Protector. This is a Data
Protector software library that is linked to the Oracle software.
Backup process
Oracle proxy-copy ZDB flow
Data Protector (10.00) Page 27 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
See the HPE Data Protector Zero Downtime Backup Administrator's Guide for a general description of the ZDB-to-disk, ZDB-to-tape, and ZDB-to-disk+tape sessions flows and for an explanation of actions triggered by ZDB options.
This section provides only the information relevant to the Data Protector Oracle ZDB integration.
Operations on a replica (mounting, activating volume/disk groups...) described below are dependent on or triggered by ZDB options. See the HPE Data Protector Zero Downtime Backup Administrator's
Guide for more information on these options.
Data Protector (10.00) Page 28 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
In the case of an offline ZDB-to-disk+tape or ZDB-to-tape session, ob2rman.pl
shuts down and opens the database instance in mount state. For both, offline and online ZDB-to-disk+tape or ZDBto-tape sessions, Data Protector starts RMAN in proxy-copy mode.
In the case of an offlineZDB-to-disk session, the database is shut down.
l
Data Protector retrieves a list of files or disk images to be included in the replica creation from the
Oracle database and starts the resolving process. The list is used only to determine the source volumes to be replicated.
In the case of a ZDB-to-disk session, if the location for control file copy is specified during configuration, Data Protector makes a copy of the control file to the specified directory on the application system. This directory has to reside on a disk array source volume.
l
In the case of an online backup, the Oracle target database is switched into the backup mode.
l ob2smbsplit or MMLstarts the procedure to create a replica of the source volumes on which the database is installed.
l
In the case of an online backup, the database files are taken out of the backup mode after the replica is created.
In the case of an offline backup, the Oracle alter database open
command is sent after the replica is created.
l
Data Protector (for ZDB to disk) or MML (for ZDB to tape or ZDB to disk+tape) starts the procedure to prepare the replica on the backup system. In this step, volume/disk groups on the backup system are enabled (UNIX systems) and, unless the database is installed on raw disks, the mount points containing the Oracle database files are mounted.
l
A ZDB agent then mounts the database on the backup system to the mount points with the same names (created by Data Protector) as on the application system.
l
If a ZDB-to-disk session is being performed, at this point the remaining ZDB options are processed and details of the session are written to the ZDB database. The session then finishes. The following steps in this description are not performed; therefore, RMAN is not given any information about the
ZDB-to-disk session.
l
MML on the application system sends a request to the Data Protector data movement agent
(DMA) on the backup system to back up the datafiles to tape.
l
The DMA reads data from the backup system and sends it to the General Media Agent to write the actual data to the backup device.
DMA’s role is also to disable the General Media Agent requests from accessing the application system. Thus, the database runs on the application system with greatly reduced performance degradation since the backup is performed on the backup system.
l
At the end of data transfer, the backup system is disabled (filesystems are dismounted on all platforms and volume/disk groups deactivated on UNIX systems) and links are re-established.
l
The recovery catalog and the control file are backed up automatically after the target database backup is finished on the backup system. However, you can disable this when creating a backup specification.
NOTE:
A replica of the archive logs is not created; therefore, the archive logs should be backed up from the application system, following the standard Data Protector Oracle archive logs backup procedure.
Data Protector (10.00) Page 29 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Configuring the integration
Prerequisites
l
It is assumed that you are familiar with the Oracle database administration and the basic Data
Protector functionality.
l
You need a license to use the Data Protector ZDB integration with Oracle. Additional licenses are required for instant recovery and for the online extension. For information on licensing, see the HPE
Data Protector Installation Guide.
l
Before you begin, ensure that you have correctly installed and configured the Oracle Server and
Data Protector client systems. See the: o
Latest support matrices at https://softwaresupport.hpe.com/manuals for an up-to-date list of supported versions, platforms, devices, and other information.
o
HPE Data Protector Installation Guide for instructions on how to install Data Protector on various architectures and how to install a Data Protector disk array integration (EMC, P9000 XP Array,
P6000 EVA Array, or NetApp Storage) with Oracle.
o
Oracle Recovery Manager User’s Guide and References for Oracle concepts and backup/recovery strategies.
o
Oracle Backup and Recovery Guide for the configuration and use of Recovery Manager, as well as for Oracle backup terminology and concepts.
o
Oracle Enterprise Manager User’s Guide for information on backup and recovery with the Oracle
Enterprise Manager, as well as information about SQL*Plus.
l
A Data Protector disk array integration (EMC, P9000 XP Array, P6000 EVA Array, or NetApp
Storage) must be correctly installed and configured. For installation, see the HPE Data Protector
Installation Guide. For configuration, see the HPE Data Protector Zero Downtime Backup
Administrator's Guide.
l
Oracle Server configurations with ASM: The disk array must support creation of replicas with cross-volume data consistency: o
If a disk array of the HPE P6000 EVA Disk Array Family is used, it must support multisnapping.
.
o
If a disk array of the HPE P9000 XP Disk Array Family is used, it must support the atomic split operation.
.
For details about which disk array models and disk array firmware revisions support creation of replicas with cross-volume data consistency, see the latest support matrices at https://softwaresupport.hpe.com/.
Data Protector (10.00) Page 30 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
The Oracle Server software must be installed on the application system and the Oracle target database must be open or mounted there.
l
From Oracle 12c onwards, the Oracle database on Microsoft Windows supports the use of an Oracle home user which is specified at the time of installation. This Oracle home user is used to run the
Windows services for Oracle home, and is similar to the Oracle user on Oracle Database on Linux.
For backup and restore in Oracle 12c database, if the Oracle integration agent and Media agent are running on the same Windows host, then to avoid shared memory allocation issues Oracle home user should be added to the Windows Backup Operator group.
For more information, see the Oracle documentation .
l
The Oracle recovery catalog database must be properly configured and open.
l
Oracle net services must be properly configured and running (on the application system) for the
Oracle target database and the recovery catalog. The net services are needed for the Data Protector
Oracle agent to be connected to the Oracle database on the application system through Oracle.
For more information about different connection options, see the Oracle Recovery Manager User’s
Guide and References.
For details on checking the prerequisites listed above, see Troubleshooting , on page 108
.
Note that the Data Protector Oracle integration uses RMAN for backup and restore. RMAN connection to a target database requires a dedicated server process. To ensure that RMAN does not
connect to a dispatcher when the target database is configured for a shared server, the net service name used by RMAN must include (SERVER_DEDICATED) in the CONNECT_DATA attribute of the connection string.
l
To successfully back up the recovery files residing in the flash recovery area, ensure that you have correctly configured the flash recovery area.
l
Oracle Real Application Clusters (RAC): Each node must have a dedicated disk for storing archive logs. Such disks must be NFS mounted on all other RAC nodes.
However, if the archive logs are not on a NFS mounted disk, you must modify the archive log
backup specification. See Backup of archive logs on RAC cannot be performed, on page 118
.
l
RAC: With Oracle version 11.2.0.2 and later, the control file must be created on a shared disk and be accessible from all RAC nodes, and the OB2_DPMCTL_SHRLOC environment variable must point to this location, from where the control file is backed up.
l
On Windows systems, when using the Oracle backup set ZDB method, set the omnirc option ZDB_
SMISA_AUTOMOUNTING on the backup system to 2 , in order to enable automatic volume mounting on the local system.
Limitations
l
The
MAXPIECESIZE
RMAN parameter option is not supported because the restore of multiple backup pieces created during a backup is not possible using the Data Protector Oracle integration.
l
In the Oracle Database 10g Release 2, for HP-UX systems, the host name length on which the
Oracle database is installed is limited to 8 characters l
Oracle OS authentication:
Oracle OS authentication is not supported in the following scenarios:
Data Protector (10.00) Page 31 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
1. For Oracle ZDB configuration.
2. For Oracle recovery catalog configuration.
3. For Oracle data guard database configuration.
l
The Oracle recovery catalog database must be used as RMAN repository for backup and restore operations. ZDB using the Oracle control file are not supported. This is set when configuring the
database. See Configuring Oracle databases, on page 41
.
l
The Oracle database identifier (DBID) must be a unique in a Data Protector cell. If you clone a database you must change the DBID.
l
Preview of zero downtime backup and instant recovery sessions is not available.
l
Using the Oracle proxy-copy ZDB method, individual tablespaces or datafiles cannot be backed up during a ZDB-to-disk or ZDB-to-disk+tape session (instant recovery enabled), only the whole database can be backed up.
l
The Oracle backup set ZDB method is supported on UNIX raw logical volumes only if these were created with LVM or VxVM.
l
When using the Oracle backup set ZDB method, you must reconfigure the Oracle integration if the
Oracle databases, on page 41 .
l
The single-host configuration (BC1, TF/1) is not supported for Oracle backup set ZDB sessions.
l
Object copying and object mirroring is not supported for ZDB to disk.
l
Recovery files residing in the flash recovery area cannot be backed up using ZDB.
l
Oracle Server configurations with ASM: The following limitations apply: o
Zero downtime backup (ZDB) is only supported with the HPE P6000 EVA Disk Array Family and the HPE P9000 XP Disk Array Family.
o
Only the backup set ZDB method is supported.
o
With the HPE P6000 EVA Disk Array Family, only configurations consisting of a single disk array unit are supported.
o
With the HPE P6000 EVA Disk Array Family, only one instance of ASM should be configured on one storage volume (LUN). If multiple instances are sharing the same LUN, instant recovery will overwrite data from all ASM instances.
o
Instant recovery is only supported with the HPE P6000 EVA Disk Array Family and the HPE
P9000 XP Disk Array Family.
l
Oracle Data Guard: Standby database is not supported for ZDB.
l
The Data Protector Oracle integration does not support non-ASCII characters in backup specification names.
Before you begin
l
Test whether the Oracle Server system and the Cell Manager communicate properly: Configure and run a Data Protector filesystem backup and restore on the Oracle Server system.
Data Protector (10.00) Page 32 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Identify the Oracle database user that will be used by Data Protector for backup. This user must have the
SYSDBA
privilege granted. For example, it could be the Oracle user sys
, which is created during database creation.
See the Oracle documentation for more information on user privileges in Oracle.
l
On Windows systems, if the Oracle target database and the Oracle recovery catalog are installed on two different systems, configure a domain user account that is a member of the Administrators group on both systems.
For supported Windows operating systems, use user impersonation. For details on setting accounts for the Inet service user impersonation, see the HPE Data Protector Help index: “Inet user impersonation”.
l
When using of the backup set method, if the Oracle database is installed on symbolic links, create these symbolic links on the backup system, too.
l
From the application system, using SQL*Plus, connect to the target database and recovery catalog by specifying the user, password, and net connect identifier. Connect to the target database as the database administrator and to the recovery catalog database as the recovery catalog owner.
Example
If the user name for the target database is system , password manager , net service name PROD , and the user name and password for the recovery catalog is rman and the net service name RMANCAT , then the commands will look like: sqlplus /nolog
SQL> connect system/manager@PROD as sysdba;
Connected.
SQL> connect rman/rman@RMANCAT;
Connected.
l
For online backup only, enable the Oracle automatic log archiving:
1. Shut down the Oracle target database instance on the application system.
2. Back up the entire database using a filesystem backup.
3. Select the location for archive logs: o
If SPFILE is used:
Execute: alter system set log_archive_dest=path_to_archive_logs SCOPE=SPFILE; o
If the init.ora
file is used:
Execute: log_archive_start=true log_archive_dest=path_to_archive_logs
The default path of the file is:
Windows systems:
ORACLE_HOME\database\initDB_NAME.ora
UNIX systems: ORACLE_HOME/dbs/initDB_NAME.ora
where
DB_NAME
is the name of the Oracle database instance.
4. Mount the target database and to enable the archive log mode, start SQL*Plus and type:
Data Protector (10.00) Page 33 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration startup mount alter database archivelog; alter database open;
Example
If the user name for the target database is system , password manager , instance name PROD , and
the user name and password for the recovery catalog is rman , then the commands will look like: sqlplus /nolog
SQL> connect system/manager@PROD as sysdba;
Connected.
SQL> startup mount;
SQL> alter database archivelog;
Statement processed.
SQL> archive log start;
Statement processed.
SQL> alter database open;
5. Back up the entire database.
Backup set method
For backup set method: l
Ensure that the Oracle software on the backup system and application system have the same directory structure. That means that ORACLE_HOME for both Oracle installations has to be identical.
l
Ensure that the following files are the same on the application system and the backup system.
Check also that the permissions are identical as on the application system: o names.ora
Default path:
ORACLE_HOME/network/admin/names.ora
o initDB_NAME.ora
Default path: ORACLE_HOME/dbs/init
DB_NAME
.ora
.
o orapwDB_NAME
Default path: ORACLE_HOME/dbs/orapwDB_NAME o admin/DB_NAME
Default path: ORACLE_BASE/admin/DB_NAME
Ensure that the Oracle net services on the application system and the backup system have the same directory structure. This can be accomplished by either NFS sharing of the files, manually copying the files from the application system to the backup system, or by using the UNIX rdist or tar commands to distribute the files from the application system.
l
Test whether the Oracle user can log in to the Oracle target database as the Oracle database administrator and to the Oracle recovery catalog database as the Oracle recovery catalog owner from the backup system:
Data Protector (10.00) Page 34 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
1. Export ORACLE_HOME , DB_NAME , and on UNIX systems also SHLIB_PATH variables.
2. Using SQL*Plus, connect to the Oracle recovery catalog database by specifying the user
(recovery catalog owner), password, and net connect identifier.
3. Connect to the Oracle target database locally using the Oracle Net software as the Oracle database administrator with the SYSDBA role.
Example
If the
DB_NAME
of the target database is
PROD
, the
DB_NAME
of the Oracle recovery catalog database is RMANCAT , and
ORACLE_HOME
is /oracle/PROD , then the commands will look like: su - ora id uid=101(ora) gid=101(dba) export DB_NAME=PROD oracle/PROD/bin/sqlplus
SQL> connect rman/rman@RMANCAT
Connected.
SQL> connect system/manager as sysdba
SQL> connect system/manager@PROD as sysdba;
Connected.
l
Test whether the user root
and the Oracle administrator (for example, the user oracle
) can connect to the target database and the recovery catalog database using the RMAN command on the backup system:
1. Log on as the Oracle database administrator to the backup system (for example, the user oracle ).
2. Execute the RMAN command and connect to the target database and the recovery catalog database.
Example
If the
DB_NAME
of the target database is
PROD
, the
DB_NAME
of the Oracle recovery catalog database is RMANCAT , and ORACLE_HOME is /oracle/PROD , then the commands will look like: su - ora id uid=101(ora) gid=101(dba) export DB_NAME=PROD rman target system/manager catalog rman/rman
Recovery Manager: Release 10.1.0.2.0 - Production
RMAN-06005: connected to target database: PROD
RMAN-06008: connected to recovery catalog database
RMAN> exit
Recovery Manager completed.
Data Protector (10.00) Page 35 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Configuring the HPE P6000 EVA Disk Array Family in ASM environments
Zero downtime backup and instant recovery with the HPE P6000 EVA Disk Array Family are supported for ASM configurations, provided that the P6000 EVA Array supports multisnapping. Additionally, the following prerequisites must be fulfilled: l
To enable zero downtime backup, the ASM-managed files that will be backed up must reside on raw disks, not on raw logical volumes.
Note that the maximum number of source disks that can be involved in multisnapping depends on the firmware revision of the P6000 EVA Array that will be used and the installed Command View
(CV) version. If the number of source disks selected for a zero downtime backup session exceeds this limitation, the session is aborted. For limitation details, see the HPE P6000 EVA Disk Array
Family documentation.
l
The autoextend feature of the Oracle Server ASM must be disabled.
l
To enable instant recovery of the Oracle ASM-managed data, set the omnirc variable SMISA_ALLOW_
ASM_IR to 1. For details, see the HPE Data Protector Help index: “omnirc options”.
Configuring the HPE P9000 XP Disk Array Family in ASM environments
In Oracle Server configurations that use Automatic Storage Management (ASM), both ZDB and instant recovery are supported with the HPE P9000 XP Disk Array Family, provided that the following prerequisites are met: l
Data files, control files, and redo log files must reside on separate storage volumes (LUNs). This configuration is needed if you plan to perform instant recovery.
l
The disk arrays must support the atomic split operation: o
Each storage volume (LUN) must be composed of a single LDEV.
o
The storage volumes on which the ASM-managed files reside must belong to consistency groups
(CTG) that have unique and non-zero IDs. Storage volumes that belong to the same CTG must be selected together in a backup specification. Otherwise, the backup session will fail. If you plan to back up datafiles, control files and log files in separate sessions, the corresponding storage volumes must reside in separate consistency groups.
Depending on your environment, set the following omnirc
options on the application system: n
SSEA_ATOMIC_SPLIT n
SSEA_ATOMIC_SPLIT_MULTIPLE_CTGROUPS n
SSEA_ATOMIC_SPLIT_MIXED_CONFIG
At least one option must be set to enable the atomic split operation. For details, see the online
Help index: “omnirc options”.
l
If the ASM instance name differs from
+ASM
, specify the correct name by setting the Data Protector omnirc option ORA_ASM_LCL_INSTANCE on the application system.
l
Additional requirements for instant recovery:
Data Protector (10.00) Page 36 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration o
In a cluster environment, on the active cluster node, set the Data Protector omnirc option ORA_
ASM_LCL_INSTANCE
to the name of the ASM instance running on that node.
o
If an Oracle ASM instance manages files of more than one database, you must reconfigure
Oracle Server to use a separate ASM disk group for each database.
o
The names of the ASM disk groups should not be changed after the backup.
o
The autoextend feature of the Oracle Server ASM must be disabled.
Cluster-aware systems
In cluster environment, if you intend to use the Data Protector CLI, set the Data Protector environment variable OB2BARHOSTNAME to the virtual server name. Set the variable on the Oracle Server system as follows:
Windows systems: set OB2BARHOSTNAME=virtual_server_name
UNIX systems: export OB2BARHOSTNAME=virtual_server_name
HP-UX with RAC: To enable instant recovery, create an HPE Serviceguard package containing only the virtual IP and the virtual hostname parameters and distribute it among the RAC nodes.
Linking Oracle Server with the Data Protector MML
To use the Data Protector Oracle integration, the Oracle Server software needs to be linked with the
Data Protector Oracle integration Media Management Library (MML) on every system on which an
Oracle instance is running.
You do not need to link Oracle Server with the Data Protector MML manually. When you start backups or restores using the Data Protector GUI or CLI, Data Protector automatically links Oracle Server with the correct platform–specific Data Protector MML. However, for testing purposes, you can override this automatic selection. You can manually specify which platform–specific Data Protector MML should be used by setting the Data Protector SBT_LIBRARY parameter. On how to set the parameter, see the util_cmd
man page. The parameter is saved in the Data Protector Oracle instance configuration file.
MML is invoked by the Oracle server when it needs to write to or read from devices using Data
Protector.
Oracle 12c CDB and PDB mode support
To backup and restore data, a new feature is introduced in Oracle 12c called Pluggable Database
(PDB). A PDB is a portable collection of schemas, schema objects, and non-schema objects that appears to an Oracle Net client as a non-container database (non-CDB). A CDB includes zero, one, or many customer created pluggable databases (PDBs). The PDB is located under container database
(CDB) in Oracle 12c.
Each CDB has the following containers:
Data Protector (10.00) Page 37 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Exactly one root
The root stores Oracle-supplied metadata and common users. An example of metadata is the source code for Oracle-supplied PL/SQL packages. A common user is a database user known in every container. The root container is named CDB$ROOT .
l
Exactly one seed PDB
The seed PDB is a system-supplied template that is used by CDB to create new PDBs. The seed
PDB is named PDB$SEED . You cannot add or modify objects in PDB$SEED .
l
Zero or more user-created PDBs
A PDB is a user-created entity that contains the data and code required for a specific set of features.
For example, a PDB can support a specific application, such as a human resources or sales application.
Backup Archive and Restore (BAR) GUI lists all the PDBs in the container database, selection of only one or more pluggable databases should be enabled.
Following are the restore scenarios: l
Restore both CDB and PDB l
Restore CDB and no PDB l
Restore one PDB and a test database
Configuring Oracle user accounts
Decide under which user accounts you want backups to run. Data Protector requires the following user accounts: l
Oracle operating system user account
For details, see Configuring Oracle operating system user accounts , below .
l
Oracle database user accounts
For details, see Configuring Oracle database user accounts, on page 40
.
Configuring Oracle operating system user accounts
For each Oracle database, Data Protector requires an operating system user account that has Oracle rights to back up the database. This user account usually belongs to the DBA user group (OSDBA
user). The user account under which the Oracle database is running has these rights. For example, to find such a user on UNIX systems, execute: ps -ef|grep ora_pmon_
DB_NAME
or ps -ef|grep ora_lgwr_
DB_NAME
Finding the Oracle user
Data Protector (10.00) Page 38 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
The following table explains how to configure users on different operating systems:
Client system
UNIX system
Description
Ensure that the Oracle user oracle from the Oracle Inventory group
( oinstall ) has been added to the Data Protector admin user group. For details on adding users, see the HPE Data Protector Help index: ”adding users”.
Add the OSDBA user account and root user account from both the application system and backup system to the Data Protector admin or operator user group. The OSDBA user on the backup system must have the same numerical user ID and group ID as the OSDBA user on the application system (for example, uid=101(ora) gid=101(dba) ).
TIP:
To find the user ID, connect to a system under this user account and execute:
#id
Windows system On Windows systems, Data Protector connects to the Oracle database using the Data Protector Inet service on the related system. By default, the service runs under the
Local System account
, which is automatically added to the Data Protector admin
user group. However, if you have restarted the Data Protector Inet service on the application system and backup system under OSDBA user accounts, you need to add the new users to the Data Protector admin or operator user group.
For information on adding users to Data Protector user groups, see the HPE Data Protector Help index:
“adding users”.
NOTE:
The OSDBA user account for the backup system needs to be added to a Data Protector user group only if you plan to use the Oracle backup set ZDB method.
Clusters
In cluster environments, ensure to add he following users to the Data Protector admin or operator user group: l
OSDBA user for all physical nodes l
OSDBA user for the virtual server (applicable for HPE Serviceguard clusters) l
UNIX systems: root user for all physical nodes
Data Protector (10.00) Page 39 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Example user configuration in a cluster environment
Configuring Oracle database user accounts
Identify or create the following Oracle database user accounts. You need to provide these user
.
Oracle database user accounts
User
Primary database user
Recovery catalog user
Description
Required to log in to the primary database.
The owner of the recovery catalog (for example, rman ). Required to log in to the catalog database. Needed if you use the recovery catalog.
If you are using Oracle 11g R2 or later, ensure that the owner of the Oracle recovery catalog: l is granted the CREATE ANY DIRECTORY and the DROP ANY
DIRECTORY system privileges, which are required to use the Data
Pump Export ( expdp
) and the Data Pump Import ( impdp
) utilities.
l has SELECT permissions on sys.v$instance view. Start SQL*Plus and type:
Data Protector (10.00) Page 40 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
User
Standby database user
Description
grant select on v_$instance to recovery_catatalog_user;
If you are using Oracle 12c or later, ensure that the owner of the Oracle recovery catalog: l is granted the EXEMPT ACCESS POLICY permissions
Required to log in to the standby database. Applicable only in Oracle Data
Guard environments. Needed to back up the standby database.
Configuring Oracle databases
Configuration of an Oracle database consists of providing Data Protector with the following data: l
Oracle Server home directory l
Login information to the target database l
Optionally, login information to the recovery catalog database l
Optionally, login information to the standby database l
Optionally, ASM-related information.
l
Backup method to be used and the related options
During the configuration, the util_oracle8.pl
command, which is started on the application system, saves the specified parameters in the Data Protector Oracle database specific configuration file on the
Cell Manager.
Ensure that the database is open during the configuration procedure and that you are able to connect to the database.
To configure an Oracle database, you can use the Data Protector GUI or the Data Protector CLI.
NOTE:
With Oracle Server configurations using ASM, to be able to perform instant recovery, the Oracle database must be configured using the Data Protector CLI. This is because the ASM-related parameters cannot be set using the Data Protector GUI. However, if you plan to perform ZDB sessions only, without instant recovery, the database can also be configured using the Data
Protector GUI.
Oracle OS authentication
To connect to the database, Data Protector can use Oracle listeners and Oracle OS authentication. For
Oracle OS authentication, during database backup user must provide username and usergroup for
Oracle OS user which can be used as Oracle OS database administrator. This user must belong do
Oracle dba/sysdba group. The user information is entered during creation of backup specification see
Creating backup specifications , on page 54
To configure backup specification using OS authentication, perform the following steps:
Data Protector (10.00) Page 41 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
1. On the Oracle host, create the OS user that will be used for Oracle OS authentication. The user must be in Oracle OS sysdba/dba group. For details how to create Oracle OS database administrator refer to Oracle documentation.
2. Add the user in Data Protector admin user group.
a. For host with Oracle agent and Windows platform add impersonation for this user.
3. Create backup specification.
a. On the Specify the application that you want to backup screen, select Specify OS user check box and enter the user name and group for created user.
b. On the Configure Oracle dialog box in primary tab select Use OS Authentication check box.
NOTE:
For Oracle RAC environments in the RAC Database Name text box enter the global database name.
Using the Data Protector GUI
Configure an Oracle database when you create the first ZDB backup specification for the database.
Start with the procedure described in Creating backup specifications , on page 54
Oracle databases, on page 41., on page 64 proceed as follows:
1. In the Configure Oracle dialog box and in the General page, specify the pathname of the Oracle
Server home directory.
Configuring Oracle - General (Windows)
Configuring Oracle - General (UNIX)
Data Protector (10.00) Page 42 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
2. In the Primary page, specify the login information to the primary database.
Note that the user must have the
SYSDBA
privilege granted.
In Services, type the net service name for the primary database instance. The backup will be performed on the system where this database instance resides.
RAC: List all net services names for the primary database separated by a comma.
Configuring Oracle - Primary
Oracle OS authentication
In the primary tab select Use OS Authentication checkbox.
Data Protector (10.00) Page 43 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
3. In the Catalog page, select Use target database control file instead of recovery catalog to use the primary database control file.
To use the recovery database catalog as an RMAN repository for backup history, select Use
recovery catalog and specify the login information to the recovery catalog.
Note that for ZDB, you must use the recovery catalog.
The user specified must be the owner of the recovery catalog.
In Services, type the net service name for the recovery catalog.
Configuring Oracle - Catalog
4. If you have Oracle Data Guard configuration for non-ZDB sessions and if you intend to back up a standby database, configure also the standby database:
In the Standby page, select Configure standby database and specify the login information to the standby database.
In Services, type the net service name for the standby database instance.
RAC: List all net services names for the standby database separated by a comma.
Configuring Oracle - Standby
5. In the ZDB page, select Backup method and then select PROXY or BACKUP SET in the dropdown list.
Data Protector (10.00) Page 44 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
In Backup control file copy location, you can specify the location on the source volumes where a backup copy of the current control file will be made during ZDB to disk.
If you do not specify the location, ob2rman.pl
will copy the copy of the control file from the application system to the backup system when it is needed. Thus, you do not need to create an additional disk for this location if you do not need the control file copy on a replica.
If your backup method is backup set and if your database instance uses PFILE (and not SPFILE), select the Parameter file (PFILE) option and specify the pathname of PFILE residing on the application system.
Configuring Oracle - ZDB
6. Click OK.
The Oracle database is configured. Exit the GUI or proceed with creating the backup specification at
Select the Oracle database objects to be backed up., on page 64
.
Using the Data Protector CLI
1. On UNIX systems, log on to the Oracle Server system with an OSDBA user account.
2. On the Oracle Server system, execute:
Windows systems:
perl -I..\lib\perl util_oracle8.pl -config -dbname DB_NAME -orahome ORACLE_HOME
[ZDB_OPTIONS] [ASM_OPTIONS] [-client CLIENT_NAME]
UNIX systems:
USEOSAUTHENTICATION [CATALOG_DB_LOGIN] [STANDBY_DB_LOGIN] [ZDB_OPTIONS] [ASM_
OPTIONS] [-client CLIENT_NAME] where:
PRIMARY_DB_LOGIN
-prmuser PRIMARY_USERNAME
-prmpasswd PRIMARY_PASSWORD
USEOSAUTHENTICATION
-useosauth USE_OS_AUT
Data Protector (10.00) Page 45 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
-racdbname RAC_DB_NAME
CATALOG_DB_LOGIN
is:
-rcuser CATALOG_USERNAME
-rcpasswd CATALOG_PASSWORD
-rcservice CATALOG_NET_SERVICE_NAME
STANDBY_DB_LOGIN
is:
-stbuser STANDBY_USERNAME
-stbpasswd STANDBY_PASSWORD
-stbservice STANDBY_NET_SERVICE_NAME_1[,STANDBY_NET_SERVICE_NAME_2 ...]
ZDB_OPTIONS
are:
-zdb_method {PROXY | BACKUP_SET}
[-ctlcp_location BACKUP_CONTROL_FILE_COPY_LOCATION]
[-pfile PARAMETER_FILE]
[-bkphost BACKUP_SYSTEM]
ASM_OPTIONS
are:
[-asmhome ASM_HOME]
[-asmuser ASM_USER -asmpasswd ASM_PASSWORD -asmservice ASM_NET_SERVICE_NAME_1[,ASM_
NET_SERVICE_NAME_2 ...]]
If you have Oracle Data Guard configuration for non-ZDB sessions and if you intend to back up a standby database, you must provide the
STANDBY_DB_LOGIN
information.
To configure an Oracle database for ZDB, you must provide the
ZDB_OPTIONS
information. If your ZDB method is backup set, you must also provide the
BACKUP_SYSTEM
information.
The
ASM_OPTIONS
options are needed for instant recovery in Oracle Server configurations that use
Automatic Storage Management (ASM).
Parameter description
CLIENT_NAME
DB_NAME
ORACLE_HOME
PRIMARY_
USERNAME
PRIMARY_
Name of the Oracle Server system with the database to be configured. It must be specified in a cluster environment or if the ZDB configuration is run on the backup system.
RAC: The virtual server of the Oracle resource group.
Oracle Data Guard: Name of either a primary system or secondary (standby) system.
Name of the database to be configured.
Pathname of the Oracle Server home directory.
Username and password for login to the target or primary database. Note that the user must have the SYSDBA privilege granted.
Data Protector (10.00) Page 46 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
PASSWORD
PRIMARY_NET_
SERVICE_
NAME_1
[,PRIMARY_
NET_SERVICE_
NAME_2, ...]
USE_OS_AUT
Net services names for the primary database.
RAC: Each net service name must resolve into a specific database instance.
This parameter is used to communicate to the configuration script to use Oracle
OS authentication instead of username and pass authentication.
Value should be 1.
RAC_DB_NAME
CATALOG_
USERNAME
CATALOG_
PASSWORD
CATALOG_NET_
SERVICE_NAME
STANDBY_
USERNAME
STANDBY_
PASSWORD
STANDBY_NET_
SERVICE_
NAME_1
[,STANDBY_
NET_SERVICE_
NAME_2, ...]
BACKUP_
CONTROL_
FILE_COPY_
LOCATION
PARAMETER_
FILE
This parameter is used in Oracle RAC environments in order to find Oracle SID for current Oracle RAC node. This parameter represent Oracle global database name and this is mandatory on Oracle RAC platforms.
Username and password for login to the recovery catalog. This is optional and is used only if you use the recovery catalog database as an RMAN repository for backup history.
Net service name for the recovery catalog.
This is used in Oracle Data Guard environment for backing up a standby database.
Username and password for login to the standby database.
Net services names for the standby database.
A location on a source volume where a copy of the current control file is made before a ZDB to disk. This is optional and if not specified, ob2rman.pl
will copy the copy of the control file from the application system to the backup system when it is needed. Thus, you do not need to create an additional disk for this location if you do not need the control file copy on a replica.
Full pathname of the PFILE residing on the application system. This is optional and used if backup method is backup set and the database instance uses PFILE (and not SPFILE).
BACKUP_
SYSTEM
ASM_HOME
Name of the backup system. It must be specified for a ZDB backup set configuration.
Home directory of the ASM instance in an Oracle ASM configuration. Specify this option if the value differs from the home directory of the Oracle database instance.
Data Protector (10.00) Page 47 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
ASM_USERNAME
ASM_PASSWORD
User name and password (authentication credentials) used by the Data Protector
Oracle integration agent to connect to the ASM database.
ASM_NET_
SERVICE_
NAME_1[,ASM_
NET_SERVICE_
NAME_2 ...]
Name of the net service to be used to access the ASM database. For Oracle environments involving multiple net services, multiple names can be specified.
The message *RETVAL*0 indicates successful configuration, even if followed by additional messages.
NOTE:
If you need to export some variables before starting SQL*Plus, listener, or RMAN, these variables must be defined in the
Environment
section of the Data Protector Oracle global configuration file or using the Data Protector GUI.
Example
The following example represents configuration on a UNIX system of an Oracle database and its recovery catalog with the backup set method used and the parameter file location specified.
The following names are used in the example: l database name: oracle l
Oracle Server home directory: /app10g/oracle10g/product/10.1.0
l primary user name: system l primary password: manager l primary net service name 1: netservice1 l primary net service name 2: netservice2 l recovery catalog user name: rman l recovery catalog password: manager l recovery catalog net service name: catservice l backup system name: bcksys
Syntax
/opt/omni/lbin/util_oracle8.pl -config -dbname oracle -orahome
/app10g/oracle10g/product/10.1.0 -prmuser system -prmpasswd manager -prmservice netservice1,netservice2 -rcuser rman -rcpasswd manager -rcservice catservice -zdb_ method BACKUP_SET -pfile /app10g/oracle10g/product/10.1.0/dbs/pfile.ora -bkphost bcksys
Example
The following example shows how to configure an Oracle database in a cluster environment. The database files are managed by ASM. The configuration enables instant recovery in ASM environments: l
Database name: SUN l
Oracle Server home directory: /orahome/ora/app/oracle/product/11.2.0/dbhome_1 l
Primary user name: sys
Data Protector (10.00) Page 48 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Primary password: oracle l
Primary net service name 1: SUN1 l
Primary net service name 2: SUN2 l
Recovery catalog user name: rman l
Recovery catalog password: manager l
Recovery catalog net service name: RECO l
ZDB method: Backup set l
Oracle Server system (cluster virtual system): cluster.company.com
l
Backup system: backup.company.com
l
ASM home directory: /oracle/crshome/crshome/crs/app/11.2.0/grid l
ASM user: sys l
ASM user password: oracle l
ASM net service name 1: ASMSRV1 l
ASM net service name 2: ASMSRV2
To configure the database, execute: opt/omni/lbin/util_oracle8.pl -config -dbname SUN -orahome
/orahome/ora/app/oracle/product/11.2.0/dbhome_1 -prmuser sys -prmpasswd oracle prmservice SUN1,SUN2 -rcuser rman -rcpasswd manager -rcservice RECO -zdb_method
BACKUP_SET -bkphost backup.company.com -client cluster.company.com -asmhome
/crshome/crs/app/11.2.0/grid –asmuser sys –asmpasswd oracle –asmservice ASMSRV1,
ASMSRV2
Checking the configuration
You can check the configuration of an Oracle database after you have created at least one backup specification for the database. If you use the Data Protector CLI, a backup specification is not needed.
Using the Data Protector GUI
1. In the Context List, select Backup.
2. In the Scoping Pane, expand Backup Specifications and then Oracle Server. Click the backup specification to display the server with the database to be checked.
3. Right-click the server and click Check configuration.
IMPORTANT:
Data Protector does not check if the specified user has appropriate Oracle backup permissions.
Using the Data Protector CLI
1. On UNIX systems, log on to the application system with an OSDBA user account.
2. Execute:
Windows systems:
Data Protector (10.00) Page 49 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration perl -I..\lib\perl util_oracle8.pl -chkconf_smb -dbname DB_NAME
UNIX systems:
util_oracle8.pl -chkconf_smb -dbname DB_NAME
Handling errors
If an error occurs, the error number is displayed in the form
*RETVAL*error_number
.
To get the error description, on the Cell Manager, execute:
Windows systems: Data_Protector_home\bin\omnigetmsg 12 error_number
UNIX systems: /opt/omni/lbin/omnigetmsg 12 error_number
IMPORTANT:
On UNIX systems, it is possible that although you receive
*RETVAL*0
, backup still fails because Data Protector does not check if the specified user has appropriate Oracle backup permissions.
Checking configuration for instant recovery
Check if the Oracle configuration is suitable for instant recovery.
On the application system, execute:
Windows systems:
perl util_oracle8.pl -chkconf_ir -dbname DB_NAME
UNIX systems:
util_oracle8.pl -chkconf_ir -dbname DB_NAME
If the control files, SPFILE, and online redo logs are on the same volume group (if LVM is used) or source volume as datafiles, a warning is displayed stating that instant recovery is not possible. You can either: l
on how to move the control files and redo logs to source volumes that are not replicated.
Or: l
Set the
ZDB_ORA_INCLUDE_CF_OLF
,
ZDB_ORA_INCLUDE_SPF
, and
ZDB_ORA_NO_CHECKCONF_
IRomnirc
options and ignore the warning. However, note that the control file, SPFILE, and online
redo logs are overwritten during instant recovery. See ZDB integrations omnirc options, on page 408
on how to set the omnirc options.
Setting environment variables
Use environment variables to modify backup environment to suit your needs. Environment variables are Oracle database specific. It means that they can be set differently for different Oracle databases.
Once specified, they are saved to related Data Protector Oracle database configuration files.
Data Protector (10.00) Page 50 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
.
NOTE:
Environment variables are not supported on HP OpenVMS systems.
Environment variables
Environment variable
OB2_RMAN_COMMAND_TIMEOUT
OB2_SQLP_SCRIPT_TIMEOUT
OB2_DPMCTL_SHRLOC
Default value
Description
300 s
300 s
This variable is applicable when Data Protector tries to connect to a target or catalog database. It specifies how long (in seconds) Data Protector waits for RMAN to respond that the connection succeeded. If RMAN does not respond within the
specified time, Data Protector aborts the current session.
This variable is applicable when Data Protector issues an SQL*Plus query. It specifies how long
Data Protector waits for SQL*Plus to respond that the query completed successfully. If
SQL*Plus does not respond within the specified time, Data Protector aborts the current session.
N/A Defines the location at which the control file is created and from where it is backed up in Data
Protector managed control file backup. Data
Protector copies the control file to its temporary files directory. This variable overrides the default directory with a customer-specified directory. In an Oracle Real Application Clusters (RAC) environments with Oracle version 11.2.0.2 or later, to enable Data Protector managed control file backups and the corresponding restore sessions, ensure this directory resides on a shared disk that all RAC nodes can access.
To set environment variables, use the Data Protector GUI or CLI.
Using the Data Protector GUI
You can set a variable when you create a backup specification or modify an existing one:
1. In the Source page of the backup specification, right-click the Oracle database at the top and click
Set Environment Variables.
2. In the Advanced dialog box, specify the variable name, its value, and click Add.
See Setting environment variables, on the next page
.
Data Protector (10.00) Page 51 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Setting environment variables
3. Click OK.
Using the Data Protector CLI
Execute: util_cmd –putopt Oracle8 DatabaseNameVariableValue –sublist Environment
For details, see the util_cmd man page or the HPE Data Protector Command Line Interface
Reference.
Example
To set the environment variable OB2_RMAN_COMMAND_TIMEOUT to 100 seconds for the Oracle database
INST2 , execute: util_cmd –putopt Oracle8 INST2 OB2_RMAN_COMMAND_TIMEOUT 100 –sublist Environment
Switching between Oracle backup methods
You can switch between the Oracle backup methods by reconfiguring the Data Protector Oracle integration for each database. It is not possible to select the method during the backup specification creation.
Data Protector (10.00) Page 52 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
IMPORTANT:
When switching between the Oracle backup set and proxy-copy methods, you must carefully follow the instructions given bellow to ensure a successful switch between both methods and to ensure that during a restore or recovery RMAN does not select backup objects backed up using different methods in one restore session. If such a mixed set is used, the restore procedure will fail.
To switch between the backup methods:
1. Successfully back up the entire database using the currently selected method.
2. To avoid selecting backup specifications with a backup method different than the current backup method, you may remove or move all ZDB backup specifications belonging to the selected database instance. The backup specifications are located on the Cell Manager in:
Windows systems:
Data_Protector_program_data\Config\Server\BarLists\Oracle8
UNIX systems: /etc/opt/omni/server/barlists/oracle8
3. Re-configure the database with the new method selected while creating a new Oracle ZDB specification.
4. Optionally, if you switch from backup set to proxy-copy, you may:
a. On the Cell Manager, remove the file:
Windows systems: Data_Protector_program_data \Config\Server\Integ\Config\
Oracle8\client_name%initDB_NAME_bckp.ora
UNIX systems: /etc/opt/omni/server/integ/config/Oracle8/ client_name%initDB_
NAME_bckp.ora
b. Remove the Oracle software from the backup system.
5. Perform ZDB of the entire database.
IMPORTANT:
If you need to perform a restore from a time between the start and the end of the first backup of the entire database using the new backup method, RMAN may try to use backup files from old method through a channel allocated for the files from the old method
and the restore will fail. See Troubleshooting , on page 108
on how to restore such a backup.
Backup
To configure an Oracle ZDB, perform the following steps:
1. Configure the devices you plan to use for a backup. For instructions, see the HPE Data Protector
Help index: “configuring devices”.
2. Configure media pools and media for a backup. For instructions, see the HPE Data Protector Help index: “creating media pools”.
3. Ensure you are able to connect to the database.
4. Configure a non-ZDB backup specification and run the backup of Oracle data on the application system to verify that you have properly configured the Oracle environment. On how to create a
Data Protector (10.00) Page 53 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration non-ZDB backup specification, see the HPE Data Protector Integration Guide for Oracle and SAP.
5. Create a Data Protector Oracle ZDB backup specification.
See Creating backup specifications , below .
Creating backup specifications
Online ZDB
To perform an online ZDB of an Oracle database, the database has to run in the ARCHIVELOG mode.
Offline ZDB
To perform an offline ZDB, create only a ZDB backup specification.
Cluster-aware systems
Before you perform an offline ZDB in a cluster environment, take the Oracle Database resource offline and bring it back online after the replica is created. This can be done using the Oracle fscmd command line interface commands in the Pre-exec and Post-exec commands for the client system in a particular backup specification, or by using the Cluster Administrator.
You cannot perform a ZDB of the archived redo log files. Therefore, you need to create two backup specifications: l
ZDB backup specification for backing up database files l standard Data Protector Oracle integration backup specification for backing up the application system archived log files
Procedure
To create an Oracle ZDB backup specification:
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click Oracle Server, and click Add
Backup.
3. In the Create New Backup dialog box, select the following:
Backup set method
To perform a ZDB of the entire database using the backup set method, select the SMB_
BackupSet_Database template.
Proxy-copy method
To perform a ZDB of the entire database using the proxy-copy method, select the SMB_Proxy_
Database template.
From the Backup type drop-down list, select Snapshot or split mirror backup, and from the Sub
type drop-down list, select the appropriate disk array agent. The agent must be installed on the
Data Protector (10.00) Page 54 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Selecting an Oracle backup template and the snapshot or split mirror backup
Click OK.
4. In Application system, select the Data Protector Oracle integration client. In a non-RAC cluster environment, select the virtual server.
RAC: Select the virtual server of the Oracle resource group.
In Backup system, select the backup system.
Select other disk array-specific backup options (see Backup , on page 53
for EMC, P9000 XP Array backup options - tbd, on page 57
for P9000 XP Array, P6000 EVA Array backup options, on page 58
for P6000 EVA Array, NetApp Storage backup options, on page 60
Storage backup options, on page 61
for EMC VNX Storage, or EMC VMAX Storage backup options, on page 62
for EMC VMAX Storage. For detailed information on the backup options, press
F1.
EMC GeoSpan specifics
In the EMC GeoSpan for Microsoft Cluster Service environment, select the backup system for the active node and specify the TimeFinder configuration.
After a failover in EMC GeoSpan for MSCS, select the backup system for the currently active node and save the backup specification.
Data Protector (10.00) Page 55 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
EMC backup options
P9000 XP Array specifics
To enable instant recovery, leave the Track the replica for instant recovery option selected. It is not possible to run instant recovery with Data Protector if this option is cleared.
NOTE: It is recommended to select the option Use the same mountpoints as on the
application system, during a ZDB+IR+ORACLE backup using a P9000 XP Array On
Linux.
Data Protector (10.00) Page 56 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
P9000 XP Array backup options - tbd
P6000 EVA Array specifics
To enable instant recovery, select the Track the replica for instant recovery option.
Data Protector (10.00) Page 57 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
P6000 EVA Array backup options
Data Protector (10.00) Page 58 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
P10000 3PAR Array specifics
P10000 3PAR Array backup options
Data Protector (10.00) Page 59 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
NetApp Storage specifics
NetApp Storage backup options
Data Protector (10.00) Page 60 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
EMC VNX Storage specifics
EMC VNX Storage backup options
Data Protector (10.00) Page 61 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
EMC VMAX Storage:
EMC VMAX Storage backup options
Click Next.
5. In Application database, type the name of the database to be backed up.
The database name can be obtained using SQL*Plus:
SQL>select name from v$database;
NOTE:
In a single-instance configuration, the database name is usually the same as its instance name. In this case, the instance name can be also used. The instance name can be obtained as follows:
SQL>select instance_name from v$instance;
Specify the User and group/domain options, which are available on UNIX and Windows Server
2008 systems, as follows: l
UNIX systems: In Username and Group/Domain name, specify the OSDBA user account under which you want the backup to start (for example, the user name ora , group DBA ). This
user must be configured as described in Configuring Oracle user accounts, on page 38
.
l
Windows Server 2008 systems: It is not mandatory to specify these options and if they are not specified, the backup runs under the Local System Account.
Data Protector (10.00) Page 62 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
In Username and Group/Domain name, specify the operating system user account under which you want the backup session to run (for example, the user name
Administrator
, domain
DP
). This user must be set up for the
Data Protector Inet
service user impersonation.
For details on setting accounts for the Inet service user impersonation, see the HPE Data
Protector Help index: “Inet user impersonation”.
Ensure that this user has been added to the Data Protector admin
or operator
user group and has the Oracle database backup rights. This user becomes the backup owner.
NOTE:
If this is not your first backup specification, Data Protector fills in Username and
Group/Domain name for you, providing the values of the last configured Oracle database.
Specifying an Oracle Server system (Windows)
Specifying an Oracle Server system (UNIX)
Data Protector (10.00) Page 63 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Click Next.
NOTE:
When you click Next, Data Protector performs a configuration check.
UNIX systems: The check is started under the specified OSDBA user account. If it completes successfully, the OSDBA user and group are also saved in both the Oracle database specific configuration file and Oracle system global configuration file, overriding previous values if they exist.
You must enter username and group name for the user for which you want to use the OS
Authentication.
6. If the Oracle database is not configured yet for use with Data Protector, the Configure Oracle dialog box is displayed. Configure the Oracle database for use with Data Protector as described in
Configuring Oracle databases, on page 41 .
7. Select the Oracle database objects to be backed up.
NOTE:
Since temporary tablespaces do not contain permanent database objects, RMAN and
Data Protector do not back them up. For more information, see the Oracle documentation.
Selecting backup objects
Data Protector (10.00) Page 64 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Click Next.
If the backup method configured for this instance does not correspond to the method in the backup specification, Data Protector will display a warning and abort the configuration.
8. Select the device(s) you want to use for the backup. Click Properties to set the device concurrency, media pool, and preallocation policy. For more information on these options, click
Help.
You can also specify whether you want to create additional copies (mirrors) of the backup during the backup session. Specify the desired number of mirrors by clicking the Add mirror and
Remove mirror buttons. Select separate devices for the backup and for each mirror.
For detailed information on the object mirror functionality, see the HPE Data Protector Help index:
“object mirroring”.
NOTE:
Object mirroring is not supported for ZDB to disk.
Click Next to proceed.
9. Set the backup options.
For information on other the Backup Specification Options and Common Application Options, press F1.
Offline ZDB
To perform an offline ZDB, select the Backup offline option in the Application Specific Options dialog box. This option stops the database before creating a replica, and restarts it after the replica
Data Protector (10.00) Page 65 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration is created. Note that if a ZDB-to-tape or ZDB-to-disk+tape session is being performed, the
database is not offline during the actual backup to tape. See Backup offline option, below .
Backup offline option
Click Next.
10. Click Save As to save the backup specification, specify a name and a backup specification group. It is recommended that you save all Oracle backup specifications in the Oracle group.
Optionally, you can click Save and Schedule to save, and then schedule the backup specification. For more information on how to create and edit schedules, see Scheduler in Data
Protector in HPE Data Protector Administrator's Guide.
Note that only the Full backup type is supported.
IMPORTANT:
The word DEFAULT is a reserved word and therefore must not be used for backup specification names or labels of any kind. Therefore, do not use a punctuation in the names
Data Protector (10.00) Page 66 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration of backup specifications, since the Oracle channel format is created from the backup specification name.
Saving the backup specification
Click OK.
To start the backup, see Starting backup sessions , on page 72 .
11. For online backup, also create a standard Data Protector Oracle integration backup specification for backing up the application system archived log files. See the HPE Data Protector Integration
Guide.
TIP:
The backup specification for the backup of archived log files can be either triggered by the
Post-Exec command defined in the ZDB backup specification for the backup of database files (recommended), or started manually after the ZDB backup specification has been started. See the HPE Data Protector Help index: “pre- and post-exec commands” for more information on configuring the Pre-Exec and Post-Exec commands.
Oracle backup options
Disable recovery catalog auto backup
Disable Data Protector managed control file backup
Back up standby database
RMAN Script
By default, Data Protector backs up the recovery catalog after every ZDB to tape or ZDB to disk+tape. Select this option to disable backup of the recovery catalog.
By default, Data Protector backs up the Data Protector managed control file after every ZDB to tape or ZDB to disk+tape. Select this option to disable backup of the Data Protector managed control file.
This option is ignored for ZDB.
You can edit the Oracle RMAN script section of the Data
Protector Oracle backup specification. The script is created by
Data Protector during the creation of a backup specification and
Data Protector (10.00) Page 67 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Pre-exec, Post-exec
Backup offline
reflects the backup specification’s selections and settings. You can edit the script only after the backup specification has been saved. For information on how to edit the RMAN script section,
see Editing the Oracle RMAN script , on the next page
.
Specify a command or RMAN script that will be started by ob2rman.pl
on the Oracle Server system before the backup
( pre-exec ) or after it ( post-exec ). RMAN scripts must have the
.rman
extension. Do not use double quotes.
For example, you can provide scripts to shut down and start up an
Oracle instance. For examples of shut-downing and starting an
Provide the pathname of the command or RMAN script.
Select this option to perform an offline ZDB session. This option stops the database before creating a replica, and restarts it after
the replica is created. See Backup offline option, on page 66
.
Examples of pre-exec and post-exec scripts on UNIX systems
Pre-exec example
The following is an example of a script that shuts down an Oracle instance:
#!/bin/sh export ORACLE_HOME=$2 export ORACLE_SQLNET_NAME=$1 if [ -f $ORACLE_HOME/bin/sqlplus ]; then
$ORACLE_HOME/bin/sqlplus << EOF connect sys/manager@$ORACLE_SQLNET_NAME as sysdba shutdown
EOF echo "Oracle database \"$DB_NAME\" shut down." exit 0 else echo "Cannot find Oracle SQLPLUS ($ORACLE_HOME/bin/sqlplus)." exit 1 fi
Post-exec example
The following is an example of a script that starts an Oracle instance:
#!/bin/sh export ORACLE_HOME=$2 export ORACLE_SQLNET_NAME=$1 if [ -f $ORACLE_HOME/bin/sqlplus ]; then
$ORACLE_HOME/bin/sqlplus << EOF
Data Protector (10.00) Page 68 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration connect sys/manager@$ORACLE_SQLNET_NAME as sysdba startup
EOF echo "Oracle database \"$DB_NAME\" started." exit 0 else echo "Cannot find Oracle SQLPLUS ($ORACLE_HOME/bin/sqlplus)." exit 1 fi
Editing the Oracle RMAN script
The RMAN script is used when the Data Protector backup specification is started to perform a backup of the Oracle objects.
The RMAN script section is not written to the backup specification until the backup specification is either saved or manually edited by clicking the Edit button.
You can edit the RMAN script section of only after the Data Protector Oracle backup specification has been saved.
Limitations
When editing the RMAN script sections of the Data Protector backup specifications, consider the following limitations: l
The Oracle manual configuration convention must be used and not the Oracle automatic configuration convention.
l
Double quotes (") must not be used - single quotes should be used instead.
l
By default, RMAN scripts created by Data Protector contain instructions for backing up one or more of the following objects: o
Databases, tablespaces, or datafiles (the first backup command) o
Archive logs (the second backup command) o
Control files (the last backup command)
The RMAN scripts with all combinations of the above listed backup objects are recognized by Data
Protector as its own scripts and it is possible to modify the selection of objects that will be backed up in the Source tab of the Results Area.
If the RMAN script contains additional manually entered backup commands, for example a second backup command for backing up a database that is already listed in the first backup command, the object selection is disabled and it is only possible to browse the Source tab.
To edit an Oracle RMAN script, click Edit in the Application Specific Options window (see
Recovery catalog settings dialog, on page 78
), edit the script, and then click Save to save the changes to the script.
See the Oracle Recovery Manager User’s Guide and References for more information on Oracle
RMAN commands.
Data Protector (10.00) Page 69 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Data Protector RMAN script structure
The RMAN script created by Data Protector consists of the following parts: l
The Oracle channel allocation together with the Oracle environment parameters’ definition for every allocated channel.
For all backup specifications except for Oracle proxy-copy ZDB backup specifications, the number of allocated channels is the same as the sum of concurrency numbers for all devices selected for backup.
NOTE:
Once the backup specification has been saved, changing the concurrency number does not change the number of allocated channels in the RMAN script. This has to be done manually by editing the RMAN script.
IMPORTANT:
On Windows systems, a maximum of 32 or 64 (if device is local) channels can be allocated.
If the calculated number exceeds this limitation, you have to manually edit the RMAN script and reduce the number of allocated channels.
When an Oracle channel is manually defined by editing the RMAN script, the environment parameters must be added in the following format: parms 'ENV=(OB2BARTYPE=Oracle8, OB2APPNAME=DB_NAME, OB2BARLIST=Backup_
Specification_Name)';
Proxy-copy
For Oracle proxy-copy ZDB backup sessions, Data Protector allocates one channel.
For Oracle proxy-copy ZDB, the OB2SMB parameter must be set to 1. If you use the Blank Oracle
Backup template, the number of concurrently running DMA ( OB2DMAP ) is automatically calculated as the sum of all device concurrences; for example, if there are 4 devices with concurrency set to 3 then OB2DMAP will be set to 12.
If you use the Oracle_SMB template, the OB2DMAP parameter is set to 1. To improve the backup and restore performance, you may want to increase the value of this parameter. The environment parameters must be added in the following format: parms 'ENV=(OB2BARTYPE=Oracle8, OB2APPNAME=DB_NAME, OB2BARLIST=Backup_
Specification_Name, OB2SMB=1, OB2DMAP=Concurrent_DMAs)';
NOTE:
The OB2DMAP parameter does not change after it has been calculated, even if you adjust the device concurrency. To change OB2DMAP , you have to manually edit the RMAN script.
l
Depending on the backup objects selection, an RMAN backup statement for the backup of the
whole database instance, and/or for any combination of RMAN commands to back up
tablespaces and datafile. The backup
statement consists of the following: o
The Oracle format of the backup file in the following format: format 'Backup_Specification_Name<DB_NAME_%s:%t:%p>.dbf' database;
NOTE:
When an Oracle format of the backup file is manually defined or changed by editing the
Data Protector (10.00) Page 70 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
RMAN script, any user-defined combination of the Oracle substitution variables can be
added to the %s:%t:%p substitution variables and
DB_NAME
, which are obligatory.
o
In case of an Oracle proxy-copy ZDB-to-disk+tape or ZDB-to-tape session, the PROXY ONLY option is required. Only one BACKUP command with the proxy only option is permitted and only one additional backup command for backing up the control file is permitted.
o
The RMAN datafile tablespace_name*datafile_name command.
l
If the archived redo logs were selected for a backup, an RMAN backup statement for the backup
of Oracle archive logs.
The backup statement consists of the Oracle format of the backup file: format 'Backup_Specification_Name<DB_NAME_%s:%t:%p>.dbf'
NOTE:
When an Oracle format of the backup file is manually defined or changed by editing the
RMAN script, any user-defined combination of the Oracle substitution variables can be
added to the obligatory %s:%t:%p substitution variables and
DB_NAME
.
l
If the control file was selected for a backup, an RMAN backup statement for the backup of
Oracle control files. The backup statement consists of the following: o
The Oracle format of the backup file in the following format: format 'Backup_Specification_Name<DB_NAME_%s:%t:%p>.dbf' current controlfile;
NOTE:
When an Oracle format of the backup file is manually defined or changed by editing the
RMAN script, any user-defined combination of the Oracle substitution variables can be
added to the %s:%t:%p substitution variables and
DB_NAME
, which are obligatory.
o
The RMAN current controlfile command.
For Oracle proxy-copy ZDB to disk or disk+tape, it is not possible to select only the control file. You must also select either a DATABASE , TABLESPACE , or DATAFILE object.
Example of the Oracle proxy-copy ZDB to disk+tape RMAN script
The following is an example of the RMAN script section as created by Data Protector based on the
Oracle SMB_Proxy_Database template, after the whole database selection: run { allocate channel 'dev_0' type 'sbt_tape' parms
'ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=DIPSI,OB2BARLIST=New1,
OB2SMB=1,OB2DMAP=1)'; backup incremental level <incr_level>format 'New1<DIPSI_%s:%t:%p>.dbf' proxy only database
; backup format 'New1<DIPSI_%s:%t:%p>.dbf' controlfile;
}
Data Protector (10.00) Page 71 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Starting backup sessions
To run a ZDB-to-disk, ZDB-to-tape, or ZDB-to-disk+tape backup session of an Oracle database, use any of the following methods:
Backup methods
l
Schedule a backup of an existing Oracle ZDB backup specification using the Data Protector
Scheduler. For more information on how to create and edit schedules, see Scheduler in Data
Protector in HPE Data Protector Administrator's Guide.
l
Start an interactive backup of an existing Oracle ZDB backup specification using the Data Protector
GUI or the Data Protector CLI. See Running an interactive backup , on the next page
.
Considerations
Before running an Oracle ZDB session, consider the following: l
It is not possible to start a ZDB, restore, or instant recovery sessions using the same source volume on the application system at the same time. A ZDB, restore, or instant recovery session must be started only after the preceding session that is using the same source volume on the application system has finished the ZDB session or restore; otherwise, the session will fail.
l
For the backup set method, if the Oracle database is installed on symbolic links, then these symbolic links have to be also created on the backup system.
l
On P9000 XP Array, if the LVM mirroring configuration is used, Data Protector displays a warning during a backup because the volume group source volumes on the application system do not have their HPE BC P9000 XP pairs assigned. This message should be ignored.
l
If the control file, SPILE, or online redo logs are on the same source volumes as the datafiles and the
Track the replica for instant recovery option is selected, the backup session will be aborted. In this case, you need to either reconfigure the database or set the ZDB_ORA_INCLUDE_CF_OLF , ZDB_
ORA_INCLUDE_SPF
, and
ZDB_ORA_NO_CHECKCONF_IRomnirc
options. See Reconfiguring an Oracle instance for instant recovery, on page 405
or ZDB integrations omnirc options, on page 408 .
Scheduling backup sessions
Scheduling a backup session means setting the time, date, and type of a backup that starts unattended
once the scheduling options are defined and saved in the backup specification. For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE Data Protector
Administrator's Guide.
To schedule an Oracle ZDB backup specification, proceed as follows:
1. In the Data Protector Manager, switch to the Backup context.
2. In the Scoping Pane, expand Backup Specifications and then Oracle Server.
3. Right-click the backup specification you want to schedule and click Edit Schedule. The
Scheduler page opens. All schedules available for this backup specification are listed in the right pane.
4. Click the schedule you want to edit, and then click the Edit icon. The Schedule wizard opens.
Data Protector (10.00) Page 72 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
5. Review the options in the Options page. and click Next. The Recurrence page opens.
Note that only the Full backup type is supported.
In the case of a ZDB-to-disk or a ZDB-to-disk+tape session, specify the Split mirror/snapshot
backup option.
NOTE:
You can run a ZDB-to-disk or a ZDB-to-disk+tape session only if the Track the replica for
instant recovery option is selected in the backup specification.
6. Set the Recurrence pattern, and click Next. The Summary page opens.
7. Review the options in the Summary page, and click Finish.
Running an interactive backup
An interactive backup can be performed any time after a backup specification has been created and saved. You can use the Data Protector GUI or CLI.
Starting a backup using the GUI
To start an interactive ZDB session of an Oracle database using the Data Protector GUI, proceed as follows:
1. In the Context List, click Backup context.
2. In the Scoping Pane, expand Backup Specifications and then Oracle Server. Right-click the backup specification you want to use and click Start Backup.
3. In the Start Backup dialog box, selectthe Network load option. For information on network load, click Help.
Note that only the Full backup type is supported.
In the case of a ZDB-to-disk or a ZDB-to-disk+tape session, specify the Split mirror/snapshot
NOTE:
You can run a ZDB-to-disk or a ZDB-to-disk+tape session only if the Track the replica for
instant recovery option is selected in the backup specification.
Data Protector (10.00) Page 73 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Selecting ZDB to disk or ZDB to disk+tape session when starting an interactive backup
Click OK.
Starting a backup using the CLI
To start an Oracle ZDB-to-tape or ZDB-to-disk+tape session using the Data Protector CLI, execute: omnib -oracle8_list Name
To start an Oracle ZDB-to-disk session using the Data Protector CLI, execute: omnib -oracle8_list Name -disk_only where
Name
is the name of the backup specification. For more information on the omnib command, see its man page or the HPE Data Protector Command Line Interface Reference.
NOTE:
It is not possible to run a ZDB-to-disk or a ZDB-to-disk+tape session if the Track the replica
for instant recovery backup option is not selected in the backup specification.
Restore
You can restore the following database objects using both the Data Protector GUI or RMAN: l
Control files l
Datafiles l
Tablespaces l
Databases l
Recovery Catalog Databases
Using the Data Protector GUI, you can also duplicate a production database. See Duplicating an
.
The following are the available methods in Data Protector for restoring database objects:
Data Protector (10.00) Page 74 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Standard restore from backup media to the application system on LAN.
See Restoring from backup media to the application system on LAN, on the next page
.
l
Instant recovery. See Instant recovery and database recovery , on page 102
.
See also Introduction, on page 15
for an overview of recovery methods depending on the backup type and type of recovery.
Microsoft Cluster Server systems
Before you start restoring a cluster-aware Oracle server, take the Oracle Database resource offline using, for example, the Cluster Administrator utility.
See Taking the Oracle resource group offline, below
.
Taking the Oracle resource group offline
Verify that you have set the Prevent Fallback option for the Oracle resource group and Do not restart for the DB_NAME.world
resource, which is an Oracle Database resource.
Checking properties
Data Protector (10.00) Page 75 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
HPE Serviceguard systems
When restoring the database from a backup performed on a virtual host, you should set
OB2BARHOSTNAME
environment variable in the RMAN script. For example: run { allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY= Path_to_Data_Protector_MML,
ENV=(OB2BARHOSTNAME=virtual.domain.com)'; restore datafile '/opt/ora10g/oradata/MAKI/example02.dbf'; release channel dev1;
}
Prerequisites
l
An instance of Oracle must be created on the system to which you want to restore or duplicate the database.
l
The database must be in the
Mount
state if the whole database is being restored, or in the
NoMount state if the control file is being restored or a database duplication is performed.
l
You must be able to connect to the database.
One way of achieving this is by configuring static service information for your Oracle listener. For details, see the Oracle documentation. You can find an example of static service information
configuration in the troubleshooting Troubleshooting , on page 108
.
l
On Windows systems, when performing a restore from backup using the Oracle backup set ZDB method, set the omnirc option ZDB_SMISA_AUTOMOUNTING on the application system to 2 , in order to enable automatic volume mounting on the local system.
Restoring from backup media to the application system on LAN
You can restore the database objects using one of the following tools within Data Protector: l
Data Protector GUI. See Restoring Oracle using the Data Protector GUI, below .
l
RMAN. See Restoring Oracle using RMAN , on page 90
.
Restoring Oracle using the Data Protector GUI
For restore, RMAN scripts are generated with necessary commands, depending on selections made in the GUI. To use additional commands, use them manually from RMAN itself. You can also use the
workaround described in How to modify the RMAN restore script, on page 120 .
Restoring database items in a disaster recovery
In a disaster recovery situation, database objects must be restored in a certain order. The following list shows you in which order database items must be restored. Under normal conditions it is possible to restore database items in any order.
Data Protector (10.00) Page 76 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
1. Restore the recovery catalog database (if it was lost)
2. Restore the control file
3. Restore the entire database or data items
Changing the database state
Before you restore any database item or you perform a duplication of a database, ensure that the database is in the correct state:
Required database states
Item to restore Database state
Control file, duplicating a database
All other items
1
NoMount (started)
Mount
To put the database into the correct state, execute: sqlplus /nolog
SQL>connect user/password@service as sysdba ;
SQL>shutdown immediate ;
To put the database into NoMount state, execute:
SQL>startup nomount ;
To put the database into
Mount
state, execute:
SQL>startup mount;
Restoring the recovery catalog database
The Oracle recovery catalog database is exported using the Oracle export utility to a binary file and backed up by Data Protector. This file has to be restored back to the disk and then imported into the
Oracle database using the Oracle import utility. Data Protector provides a facility to do this automatically using the Oracle integration.
To restore the recovery catalog database:
1. Ensure that the recovery catalog database is in the Open state.
2. Remove the recovery catalog from the database (if it exists), using the RMAN command DROP
CATALOG .
3. In the Data Protector GUI, switch to the Restore context.
4. Under Restore Objects, expand Oracle Server, expand the system on which the database, for which you want to restore the recovery catalog, resides, and then click the database.
1
When restoring only a few tablespaces or datafiles, then the database can be open with the tablespaces or datafiles to be restored offline.
Data Protector (10.00) Page 77 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
5. In the Restore action drop-down list, select Perform RMAN Repository Restore.
In the Results Area, select RECOVERY CATALOG.
If you want to change the recovery catalog login information, right-click RECOVERY CATALOG and click Properties. In Recovery Catalog Settings, specify the login information for recovery catalog.
Recovery catalog settings dialog
6. In the Options page:
In User name and User group, specify the user name and password to the recovery catalog database.
From the Session ID drop-down list, select the Session ID.
For further information, see Restore, recovery, and duplicate options , on page 86
.
7. Click Restore.
Proceed to restore the control file.
Data Protector (10.00) Page 78 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Restoring the control file
The control file contains all the information about the database structure. If the control file has been lost, you must restore it before you restore any other part of the database. The database should be in the NoMount state.
Depending on the type of the control file backup, the following types of restore are possible when restoring the control file: l
Restoring from Data Protector managed control file backup ( CONTROLFILE FROM DP MANAGED
BACKUP
)
The control file was backed up automatically by ob2rman.pl
at the end of a backup session, unless the option Disable Data Protector managed control file backup was selected.
The recovery catalog is not required for this restore option.
The control files ( ctrlDB_NAME.dbf
) are restored to the default Data Protector temporary files directory.
NOTE:
In Oracle Real Application Clusters (RAC) environments with Oracle versions 11.2.0.2 and later, the control files are created at, backed up from, and restored to the location defined by the
OB2_DPMCTL_SHRLOC
variable. This directory must reside on a shared disk and be accessible from all RAC nodes in order for restore sessions to succeed.
After the restore, execute the following script: run { allocate channel 'dev0' type disk; restore controlfile from ' TMP_FILENAME'; release channel 'dev0';
}
Where
TMP_FILENAME
is the location to which the file was restored.
l
Restoring from RMAN backup set ( CONTROLFILE FROM RMAN BACKUPSET )
The recovery catalog is required.
A backup session can contain more than one type of the control file backup.
To restore the control file:
1. Open the sqlplus
window and put the database in the nomount state. See Changing the database state , on page 77 .
2. In the Data Protector GUI, switch to the Restore context.
3. Under Restore Objects, expand Oracle Server, expand the system on which the database, for which you want to restore the control file, resides, and then click the database.
4. In the Restore Action drop-down list, select Perform RMAN Repository Restore.
In the Results area, select the control file for restore.
5. In the Options page, from the Client drop-down list, select the system on which the Data
Protector Oracle integration agent ( ob2rman.pl
) will be started. To restore the control file to a different database than it is selected, click Settings and specify the login information for the target
Data Protector (10.00) Page 79 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
database.
6. Click Restore.
Proceed with restoring the Oracle database objects.
Restoring Oracle database objects
Before you restore Oracle database objects, ensure that you have an up-to-date version of the recovery catalog database and the control file. They contain the database structure information. If you do not
and Restoring the control file , on the previous page .
To restore Oracle database objects:
1. Put the database in the mount state. See Changing the database state , on page 77 .
2. In the Data Protector GUI, switch to the Restore context.
3. Under Restore Objects, expand Oracle Server, expand the system on which the database, for which you restore the database objects, resides, and then click the database.
4. In the Restore action drop-down list, select the type of restore you wish to perform. For
information on the options, see Restore, recovery, and duplicate options , on page 86
.
IMPORTANT:
If you do not select Perform Restore and Recovery or Perform Recovery Only, you will
have to recover the database objects manually using RMAN. For information, see Restoring
Oracle using RMAN , on page 90
.
Data Protector (10.00) Page 80 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Source page
5. In the Results Area, select objects for restore.
If you are restoring datafiles, you can restore the files to a new location. Right-click the database object, click Restore As, and in the Restore As dialog box, specify the new datafile location.
NOTE:
When restoring to a new location, current datafiles will be switched to the restored datafile copies only if you have selected Perform Restore and Recovery from the Restore
action drop-down list.
6. In the Options page, from the Client drop-down list, select the system on which the Data
Protector Oracle integration agent will be started. To restore the database objects to a different database than it is selected, click Settings and specify the login information for the target database.
Options page
Data Protector (10.00) Page 81 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
7. In the Devices page, select the devices to be used for the restore.
For more information on how to specify devices for a restore, see the HPE Data Protector Help index: “restore, selecting devices for”.
Devices page
Data Protector (10.00) Page 82 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
8. Click Restore.
After the restore:
1. Put the database in the correct state.
If you selected Perform Restore and Recovery or Perform Recovery Only in the Source page, then the database is automatically put into Open state by Data Protector.
2. If you performed an Oracle database restore and recovery until point in time, and the session has finished successfully, reset the database to register the new incarnation of database in the recovery catalog.
Connect to the target and recovery catalog database using RMAN and reset the database: rman target Target_Database_Login catalog Recovery_Catalog_Login
RMAN> RESET DATABASE;
RMAN> exit
3. If you did not choose to use Data Protector to recover the database objects and if you have all archived redo logs on disk, perform the following after the database is restored:
Open a command line window and enter the following commands: sqlplus /nolog
SQL>recover database;
SQL>connect user/password@service as sysdba;
SQL>alter database open;
Data Protector (10.00) Page 83 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Restoring tablespaces and datafiles
To restore tablespaces and datafiles:
1. Open a command line window and enter the following commands if you have the database in the
Open state: sqlplus /nolog
SQL>connect user/password@service as sysdba
;
SQL>alter database datafile 'datafile name' offline;
If you are restoring a tablespace enter:
SQL>alter tablespace tablespace_name offline;
2. When the restore has been completed put the datafiles and tablespaces back online with the following procedures:
Open a command line window and enter the following commands: sqlplus /nolog
SQL>connect user/password@service as sysdba
If you are restoring a datafile enter:
SQL>alter database datafile 'datafile_name' online;
If you are restoring a tablespace enter:
SQL>alter tablespace tablespace_name online;
Duplicating an Oracle database
Perform a production database duplication to create: l
A standby database which has the same DBID as the production (primary) database. With this, you can: o
Create a new standby database.
o
Re-create a standby database after: n
Loss of entire standby database n
Primary database control file was restored or recreated n
Database point-in-time recovery was performed on the primary database n
Switchover or failover of database roles occurred l
An independent copy, with a unique DBID, which can be used for data mining or testing purposes.
Prerequisites
l
The whole primary database with the archived logs must be backed up. l
Archive logs, which have not been backed up to tape since the last full backup and are required for duplication must be available on the duplicate system with the same path names as on the target system (system with the production database to be duplicated).
l
Net service name for the auxiliary instance must be configured.
Data Protector (10.00) Page 84 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
When duplicating a database on the same system on which the target database resides, set all *_
PATH
,
*_DEST
,
DB_FILE_NAME_CONVERT
, and
LOG_FILE_NAME_CONVERT
initialization parameters appropriately. Thus, the target database files will not be overwritten by the duplicate database files.
Limitations
l
Database duplication is not supported using proxy copy backups of the primary database.
l
If you perform duplication of a database (not for standby) on the same system on which the target or production database resides, note that you cannot use the same database name for the target and duplicate databases when the duplicate database resides in the same Oracle home directory as the target database. Note also that if the duplicate database resides in a different Oracle home directory than the target database, then the duplicate database name has to differ from other database names in that same Oracle home directory.
To duplicate a production database:
1. On the system where the selected database will be duplicated, put the Oracle auxiliary database
instance in the nomount state. See Changing the database state , on page 77 .
2. In the Context List of the Data Protector GUI, click Restore.
3. Under Restore Objects, expand Oracle Server, expand the system on which the production database resides, and then click the production database which you want to duplicate. If there are several such systems, select the system on which you want the Data Protector Oracle integration agent ( ob2rman.pl
) to be started.
4. In the Restore Action drop-down list, select Perform Duplication.
5. In the Options page, from the Client drop-down list, select the system on which the Data
Protector Oracle integration agent ( ob2rman.pl
) will be started.
Click Settings to specify the login information (a user name, password, and net services name) for the auxiliary database. If you do not provide the login information, the duplication session will fail.
In User name and User group, specify the user name and group for the OSDBA account, which will be used by the Data Protector Oracle integration agent.
In Parallelism, specify the number of RMAN auxiliary channels to be allocated for database duplication.
Set duplicate options. For information, see Duplicate options, on page 88 or press F1.
If you are creating a new database copy (not for standby), specify also the Recover until option to recover the duplicated database until a specified point in time.
Data Protector (10.00) Page 85 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Oracle duplicate options
6. Click Restore.
When the standby database is created, it is left mounted. Start the managed recovery process (log apply services) manually.
For information on how to use the RMAN commands to duplicate a database, see the Oracle documentation.
Restore, recovery, and duplicate options
Restore action options
The following describes each of the options in the Source page. This page is used to define the combination of restore and recovery you would like to perform using the GUI.
In the context of Data Protector, “restore” means to restore the datafiles. You can select which database, tablespace, or datafiles they would like to restore and up to which point in time they would
Data Protector (10.00) Page 86 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration like them to be restored. “Recover” means applying the redo logs. You can select which redo logs to apply according to SCN number, logseq, or you can apply all the redo logs to the time of the last backup.
Perform
Restore
Use this option to only restore (but not recover) the database objects using Data
Protector. After restore, recover the database manually using RMAN. For information
on recovering the database using RMAN, see Restoring Oracle using RMAN , on page 90 .
Use this option to perform both the restore and recovery of the database objects using
Data Protector.
Perform
Restore and
Recovery
Perform
Recovery
Only
Use this option to only recover the database. This action can only be performed on the whole database.
Perform
RMAN
Repository
Restore
Use this option to restore the recovery catalog or the control file when the database objects are not available in the Source page.
Perform
Duplication
This option is used to perform duplication of a production database. This action can only be performed on the whole database.
General options
Client
Settings
This option specifies the system on which the Data Protector Oracle integration agent
( ob2rman.pl
) will be started.
Click Settings to specify the login information (user name, password, and net service name) for the target database (in case of restore and recovery) or auxiliary database
(in case of duplication) where you want the selected database objects to be restored or duplicated.
If this is not specified in the case of restore or recovery, the login information of the selected database that resides on the selected system will be used.
If this is not specified in the case of duplication, the duplication session will fail.
User name,
User group
(UNIX systems only)
Specify the operating system user account under which you want the restore to start.
Ensure that this user has Oracle rights to restore the database (for example, it is in the
DBA user group). The user must also be in the Data Protector admin or operator user group (actually, the Start restore and See private objects user rights suffice).
Restore mode
This drop-down list allows you to specify which type of restore you would like perform.
The options are: l
Normal
Data Protector (10.00) Page 87 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
This option should be used when a conventional backup or ZDB using the backup set method was performed.
l
Proxy copy
This option should be used when the original Oracle backup was made using the
Oracle RMAN proxy-copy method.
This option is disabled when you perform recovery only.
Parallelism
This field is used to specify the number of concurrent data streams that can read from the backup device. The default value is one.
In case of Normal restore mode, to optimize restore performance, specify the same number of data streams as were used during the backup. For example, if you set the backup concurrency to 3, set the number of parallel data streams to 3 as well. Note that if a very high number of parallel data streams is specified this may result in a resource problem because too much memory is being used.
For Oracle proxy-copy ZDB sessions, this option is disabled and Data Protector sets the number of concurrent data streams to the value that was used at backup. If you are restoring a backup created using a previous version of Data Protector, parallelism is set to the number of devices that were used for backup, regardless of the concurrency numbers for these devices.
Duplicate options
Available if Perform Duplication was selected.
For Standby
DORECOVER
To database name
Select this option to create a standby database.
Default: selected.
Available if For Standby was selected.
Select this option if you want RMAN to recover the database after creating it.
Select this option to create a new database copy. In the text box, specify its name. The name should match the name in the initialization parameter file that was used to start the auxiliary database instance. By default, the database name is set to the database name of the currently selected target database.
NOFILENAMECHECK
Select this option to disable RMAN to check whether the target datafiles share the same names with the duplicated datafiles.
Select this option when the target datafiles and duplicated datafiles have the same names, but reside on different systems.
Default: not selected.
Data Protector (10.00) Page 88 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Restore and recovery options
Restore until
The options in this drop-down list allow you to limit the selection to those backups that are suitable for an incomplete recovery to the specified time.
l
Now
Use this option to restore the full backup. By default, this option is selected.
l
Selected time
Use this option to specify an exact time to which you wish the database to be restored. Data Protector restores the backup that can be used in recovery to the specified time.
l
Selected logseq/thread number
A logseq number is a redo log sequence number. Use this option to specify a particular redo log sequence and a thread number which will act as an upper limit of redo logs to restore. Data Protector restores the backup that can be used in recovery to the specified log sequence number.
l
Selected SCN number
Use this option to specify the SCN number to which you wish the database to be restored. Data Protector restores the backup that can be used in recovery to the specified SCN number.
Recover until
The options in this drop-down list allow you to specify to which point in time you would like the recovery to be performed.
l
Now
Data Protector starts RMAN to recover the database to the most recent time possible by applying all archived redo logs. By default, this option is selected.
l
Selected time
Use this option to specify an exact time to which the archive logs are applied.
l
Selected logseq/thread numbe r
A logseq number is a redo log sequence number. Use this option to specify a particular redo log sequence and a thread number which will act as an upper limit of redo logs to recover.
l
Selected SCN number
Use this option to specify the SCN number to which you perform the recovery.
If you reset the logs, also reset the database; otherwise, Oracle will during the next backup try to use the logs that were already reset and the backup will fail. Login to the target and recovery catalog database and execute: rman target Target_Database_Login catalog Recovery_Catalog_Login
RMAN> RESET DATABASE;
RMAN> exit
Open database
Opens the database after a recovery is performed.
Data Protector (10.00) Page 89 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
after recovery
Reset logs
Resets the archive logs after the database is opened.
Always reset the logs: l
After an incomplete recovery (not Recover until now).
l
If a backup of a control file is used in recovery or restore and recovery.
Do not reset the logs: l
After a complete recovery (Recover until now) when the backup of a control file was not used in recovery or restore and recovery. l
On the primary database, if the archive logs are used for a standby database.
However, if you must reset the archive logs, you will need to recreate the standby database.
If you reset the logs when the Recover until option is set to Now, a warning is displayed, stating that you should reset the logs only if you use an older control file for restore.
NOTE:
Oracle recommends that you perform a complete backup immediately after a database was opened with the Reset Logs option.
Restoring Oracle using RMAN
Data Protector acts as a media management software for the Oracle system, therefore RMAN can be used for a restore.
This section only describes examples of how you can perform a restore. The examples provided do not apply to all situations where a restore is needed.
See the Oracle Recovery Manager User’s Guide and References for detailed information on how to perform: l
Restore and recovery of the database, tablespace, control file, and datafile.
l
Duplication of a database.
The following examples of restore are given: l
Example of full database restore and recovery, on page 93
l
Example of point-in-time restore, on page 94
l
Example of tablespace restore and recovery, on page 95
l
Example of datafile restore and recovery, on page 97
l
Example of archive log restore, on page 100
The restore and recovery procedure of Oracle control files is a very delicate operation, which depends on the version of the Oracle database you are using. For detailed steps on how to perform the restore of control files, see the Recovery Manager User’s Guide and References.
Data Protector (10.00) Page 90 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Preparing the Oracle database for restore
The restore of an Oracle database can be performed when the database is in mount mode. However, when you are performing the restore of tablespaces or datafiles, only a part of the Oracle database can be put offline.
Prerequisites
The following requirements must be met before you start a restore of an Oracle database: l
Make sure that the recovery catalog database is open. If the recovery catalog database cannot be
for details of how to restore the recovery catalog database.
l
Check which ZDB method (proxy-copy or backup set) was used for the backup session that you plan to restore.
l
Control files must be available. If the control files are not available, you must restore them. See the
Oracle Recovery Manager User’s Guide and References for more details.
If you have to perform a restore of the recovery catalog database, you must perform this restore first.
Only then can you perform a restore of other parts of the Oracle database.
When you are sure that the recovery catalog database files are in place, start the recovery catalog database.
l
Make sure that the following environment variables are set: o
ORACLE_BASE o
ORACLE_HOME o
ORACLE_TERM o
DB NAME o
PATH o
NLS_LANG o
NLS_DATE_FORMAT
Windows systems example
ORACLE_BASE=Oracle_home
ORACLE_HOME=Oracle_home\product\10.1.0
ORACLE_TERM=HP
DB_NAME=PROD
PATH=$PATH:Oracle_home\product\10.1.0\bin
NLS_LANG=american
NLS_DATE_FORMAT='Mon DD YYYY HH24:MI:SS'
UNIX systems example
ORACLE_BASE=/opt/oracle
Data Protector (10.00) Page 91 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
ORACLE_HOME=/opt/oracle/product/10.1.0
ORACLE_TERM=HP
DB_NAME=PROD
PATH=$PATH:/opt/oracle/product/10.1.0/bin
NLS_LANG=american
NLS_DATE_FORMAT='Mon DD YYYY HH24:MI:SS' l
Check that the /etc/oratab file has the following line:
Windows systems:
PROD:Oracle_home\product\10.1.0:N
UNIX systems: PROD:/opt/oracle/product/10.1.0:N
The last letter determines whether the database will automatically start upon boot-up (Y) or not (N).
Connection strings used in the examples
In the examples below, the following connection strings are used: l
Target connection string for target database: sys/manager@PROD where sys is the username, manager is the password and PROD is a net service name.
l
Recovery catalog connection string for recovery catalog database: rman/rman@CATAL where rman is the username and password and CATAL is a net service name.
SBT_LIBRARY parameter
On Windows and UNIX systems, set the SBT_LIBRARY RMAN script parameter to point to the correct platform-specific Data Protector MML. The parameter must be specified for each RMAN channel separately. For details on the Data Protector MML location, see the HPE Data Protector Integration
Guide.
The SBT_LIBRARY path parameter must contain no spaces such that RMAN can load the library successfully. On a Windows system, you may run into problems configuring your Oracle backup, and the configuration may fail with the following error messages:
SBT_LIBRARY=C:/Program Files/OmniBack/bin/orasbt.dll
[...]
ORA-19554: error allocating device, device type: SBT_TAPE, device name:
ORA-27209: syntax error in device PARMS - unknown keyword or missing =
If, at this stage, SBT_LIBRARY parameter contains a path with space, you must check if the path
C:\Program Files or the Data Protector installation directory,
DP_HOME
contains a shortpath (8dot3) notation with a CMD command
<dir /X>
.
If the shortpath does not exist or has been deleted, or disabled during installation, you can manually add using the following command:
<windows\System32\fsutil.exe file setshortname 'C:\Program Files' PROGRA~1>
Data Protector (10.00) Page 92 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Note that since
Program Files
directory is in use when the OS is active, the command will only work when you run the Windows repair mode CMD on your Oracle machine.
If you are not able to add the shortname because you can not take your production database offline for repair mode, use the following workaround:
1. Create a symbolic link to <DP_HOME>\bin\orasbt.dll
using CMD <mklink> and save it to a path with no spaces.
2. Set the Oracle variable SBT_LIBRARY as described in the above procedure, and the util_cmd CLI reference pages.
If there are no Oracle configurations because of new installation, you must specify optional parameter to ensure that the config file is created:
-local
"C:\ProgramData\OmniBack\Config\Server\Integ\Config\Oracle8\hostname.domain.suffix%
SID"
CAUTION:
If you use the above workaround, you are responsible for the location of the symbolic link, security, and access rights.
In the following examples, the SBT_LIBRARY parameter is set to /opt/omni/lib/libob2oracle8.so
, which is the correct path for 32-bit Solaris systems.
Example of full database restore and recovery
To perform a full database restore and recovery, you also need to restore and apply all the archive logs.
To perform a full database restore and recovery:
1. Log in to the Oracle RMAN:
Windows systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL
2. Start the full database restore and recovery:
For a non-ZDB or ZDB backup set session: run{ allocate channel 'dev1' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME)'; restore database; recover database; sql 'alter database open'; release channel 'dev1';
}
For a ZDB proxy-copy session: run{ allocate channel 'dev1' type 'sbt_tape' parms
Data Protector (10.00) Page 93 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2PROXYCOPY=1, OB2APPNAME= DB_NAME)'; restore database; recover database; sql 'alter database open'; release channel 'dev1';
}
You can also save the script into a file and perform a full database restore using the saved files. The procedure in such cases is as follows:
1. Create a restore_datafile file in the default Data Protector temporary files directory.
2. Start the full database restore:
Windows systems:
ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=Data_Protector_home\tmp\restore_datafile
UNIX systems:
ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=/var/opt/omni/tmp/restore_datafile
Example of point-in-time restore
To perform a point-in-time restore, you also need to restore and apply the archive logs to the specified point in time. To perform a point-in-time database restore and recovery:
1. Log in to the Oracle RMAN:
Windows systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL
UNIX systems:
ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL
2. Start the point-in-time restore:
For a non-ZDB or ZDB backup set session: run{ allocate channel 'dev1' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; set until time 'Mar 14 2004 11:40:00'; restore database; recover database; sql 'alter database open'; release channel 'dev1';
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for database recovery. Release the proxy-copy channel before the recovery: run{ allocate channel 'dev1' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2PROXYCOPY=1, OB2APPNAME=
DB_NAME)';
Data Protector (10.00) Page 94 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration allocate channel 'dev2' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8, OB2APPNAME= DB_NAME)'; set until time 'Mar 14 2006 11:40:00'; restore database; release channel 'dev1'; recover database; sql 'alter database open'; release channel 'dev2';
}
3. After you have performed a point-in-time restore, reset the database in the Recovery Catalog.
You can also save the script into a file and perform a point-in-time restore using the saved files:
1. Create a restore_PIT file in the default Data Protector temporary files directory.
2. Start the point-in-time restore:
Windows systems:
ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=Data_Protector_home\tmp\restore_PIT
UNIX systems:
ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=/var/opt/omni/tmp/restore_PIT
Example of tablespace restore and recovery
If a table is missing or corrupted, you need to perform a restore and recovery of the entire tablespace.
To restore a tablespace, you may take only a part of the database offline, so that the database does not have to be in the mount mode. You can use either a recovery catalog database or control files to perform a tablespace restore and recovery. Follow the steps below:
1. Log in to the Oracle RMAN:
Windows systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL
2. Start the tablespace restore and recovery.
l
If the database is in the open state, the script to restore and recover the tablespace should have the following format:
For a non-ZDB or ZDB backup set session: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; sql 'alter tablespace TEMP offline immediate'; restore tablespace TEMP; recover tablespace TEMP; sql 'alter tablespace TEMP online'; release channel dev1;
}
Data Protector (10.00) Page 95 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for database recovery. Release the proxy-copy channel before the recovery: run{ allocate channel 'dev1' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2PROXYCOPY=1, OB2APPNAME= DB_NAME)'; allocate channel 'dev2' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8, OB2APPNAME=
DB_NAME)'; sql 'alter tablespace TEMP offline immediate'; restore tablespace TEMP; release channel 'dev1'; recover tablespace TEMP; sql 'alter tablespace TEMP online'; release channel 'dev2';
} l
If the database is in the mount state, the script to restore and recover the tablespace should have the following format:
For a non-ZDB or ZDB backup set session: run{ allocate channel
dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; restore tablespace 'TEMP'; recover tablespace 'TEMP'; release channel
dev1;
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for database recovery. Release the proxy-copy channel before the recovery: run{ allocate channel 'dev1' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2PROXYCOPY=1, OB2APPNAME=
DB_NAME)'; allocate channel 'dev2' type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8, OB2APPNAME= DB_NAME)'; restore tablespace 'TEMP'; release channel 'dev1'; recover tablespace 'TEMP'; release channel 'dev2';
}
You can also save the script into a file and perform a tablespace restore using the saved files:
1. Create a restore_TAB
file in the default Data Protector temporary files directory.
2. Start the tablespace restore.
Data Protector (10.00) Page 96 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Windows systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=Data_Protector_home\tmp\restore_TAB
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=/var/opt/omni/tmp/restore_TAB
Example of datafile restore and recovery
To restore and recover a datafile, you may take only a part of the database offline.
To restore and recover a datafile:
1. Log in to the Oracle RMAN.
Windows systems:
ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL
2. Start the datafile restore and recovery: l
If the database is in an open state, the script to restore the datafile should have the following format:
UNIX systems
For a non-ZDB or ZDB backup set session: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; sql "alter database datafile
''/opt/oracle/data/oradata/DATA/temp01.dbf'' offline"; restore datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; recover datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; sql "alter database datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf' online"; release channel dev1;
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for database recovery. Release the proxy-copy channel before the recovery: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; allocate channel dev2 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME, OB2PROXYCOPY=1)'; sql "alter database datafile
''/opt/oracle/data/oradata/DATA/temp01.dbf'' offline";
Data Protector (10.00) Page 97 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration restore datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; release channel dev2; recover datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; sql "alter database datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf' online"; release channel dev1;
}
Windows systems
For a non-ZDB or ZDB backup set session: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY= Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; sql "alter database datafile
''C:\oracle\data\oradata\DATA\temp01.dbf'' offline"; restore datafile
'C:\oracle\data\oradata\DATA\temp01.dbf'; recover datafile
'C:\oracle\data\oradata\DATA\temp01.dbf'; sql "alter database datafile
''C:\oracle\data\oradata\DATA\temp01.dbf'' online"; release channel dev1;
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for the recovery process. Release the proxy-copy channel before the recovery: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY= Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME)'; allocate channel dev2 type 'sbt_tape' parms
'SBT_LIBRARY=
Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME, OB2PROXYCOPY=1)'; sql "alter database datafile
'' Oracle_home\data\oradata\DATA\temp01.dbf'' offline"; restore datafile
' Oracle_home\data\oradata\DATA\temp01.dbf'; release channel dev2; recover datafile
'
Oracle_home\data\oradata\DATA\temp01.dbf'; sql "alter database datafile
'' Oracle_home\data\oradata\DATA\temp01.dbf'' online"; release channel dev1;
} l
If the database is in a mount state, the script to restore and recover the datafile should have the
Data Protector (10.00) Page 98 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration following format:
UNIX system
For a non-ZDB or ZDB backup set session: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; restore datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; recover datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; release channel dev1;
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for the recovery process. Release the proxy-copy channel before the recovery: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; allocate channel dev2 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME, OB2PROXYCOPY=1)'; restore datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; release channel dev2; recover datafile
'/opt/oracle/data/oradata/DATA/temp01.dbf'; release channel dev1;
}
Windows system
For a non-ZDB or ZDB backup set session: run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY= Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)'; restore datafile
' Oracle_home\data\oradata\DATA\temp01.dbf'; recover datafile
'
Oracle_home\data\oradata\DATA\temp01.dbf'; release channel dev1;
}
For a ZDB proxy-copy session, allocate one channel for restoring proxy-copy sessions and one channel for the recovery process. Release the proxy-copy channel before the recovery:
Data Protector (10.00) Page 99 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration run{ allocate channel dev1 type 'sbt_tape' parms
'SBT_LIBRARY= Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME)'; allocate channel dev2 type 'sbt_tape' parms
'SBT_LIBRARY=
Data_Protector_home\bin\orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME, OB2PROXYCOPY=1)'; restore datafile
' Oracle_home\data\oradata\DATA\temp01.dbf'; release channel dev2; recover datafile
'
Oracle_home\data\oradata\DATA\temp01.dbf'; release channel dev1;
}
You can also save the script into a file and perform a datafile restore using the saved files:
1. Create a restore_dbf
file in the default Data Protector temporary files directory.
2. Start the datafile restore:
Windows systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=/var/opt/omni/tmp/restore_dbf
UNIX systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=Data_Protector_home\tmp\restore_dbf
Example of archive log restore
To restore an archive log:
1. Log in to the Oracle RMAN:
Windows systems: ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL
2. Start the archive log restore: run{ allocate channel
dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME)'; restore archivelog all; release channel dev1;}
You can also save the script into a file and perform an archive log restore using the saved files:
1. Create a restore_arch file in the default Data Protector temporary files directory.
2. Start the archive log restore:
Windows systems:
ORACLE_HOME\bin\rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=Data_Protector_home\tmp\restore_arch
Data Protector (10.00) Page 100 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@PROD catalog rman/rman@CATAL cmdfile=/var/opt/omni/tmp/restore_arch
Example of database restore using a different device (with the automatic device selection functionality disabled)
Suppose a database was backed up with the device dev1
. To restore the database with the device dev2 , add the line send device type 'sbt_tape' 'CHDEV=dev1>dev2'; to the RMAN script:
1. Log in to the Oracle RMAN:
Windows systems:
ORACLE_HOME\bin\rman target sys/manager@TIN
UNIX systems: ORACLE_HOME/bin/rman target sys/manager@TIN
2. Execute: run { allocate channel 'dev_0' type 'sbt_tape'
parms 'SBT_LIBRARY=C:/PROGRA~1/OmniBack/bin/orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=CAN,OB2BARLIST=test)'; allocate channel 'dev_1' type 'sbt_tape'
parms 'SBT_LIBRARY=C:/PROGRA~1/OmniBack/bin/orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=CAN,OB2BARLIST=test)'; allocate channel 'dev_2' type 'sbt_tape'
parms 'SBT_LIBRARY=C:/PROGRA~1/OmniBack/bin/orasbt.dll,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=CAN,OB2BARLIST=test)'; send device type 'sbt_tape' 'NO_AUTO_DEVICE_SELECTION=1'; send device type 'sbt_tape' 'CHDEV=dev1>dev2'; restore database;
}
NOTE:
l
The line device type 'sbt_tape' 'NO_AUTO_DEVICE_SELECTION=1';
disables the automatic device selection.
l
You can also specify multiple device redirections using
CHDEV
parameter with
'CHDEV=dev1>dev2;dev3>dev4'; syntax.
Restoring using another device
Data Protector supports the restore of Oracle database objects from devices other than those on which the database objects were backed up.
Specify these devices in the /etc/opt/omni/server/cell/restoredev (UNIX systems) or
Data_
Protector_program_data\Config\server\Cell\restoredev (Windows systems) file in the following format:
" DEV 1 " " DEV 2 " where
DEV 1 is the original device and DEV 2 the new device.
Data Protector (10.00) Page 101 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
On Windows systems, this file must be in the Unicode format.
Note that this file should be deleted after it is used.
Example
Suppose you have Oracle objects backed up on a device called DAT1 . To restore them from a device named DAT2 , specify the following in the restoredev file:
" DAT1 " " DAT2 "
Restoring Oracle to different cell/client
To restore Oracle to different cell/client perform the following steps:
1. Copy the Data Protector Oracle configuration files from source to target.
2. Export the Oracle SID.
3. Create password file for the database.
4. Set the DBID and start the database in nomount state.
5. Recover the pfile from RMAN.
6. Start the database in nomount state with recovered pfile.
7. Start the control file restore from GUI and copy the Data Protector restored control file to control file location.
8. Start the database in mount state.
9. Configure the database backup with control file from target side.
10. Start the database restore only.
11. Execute the sql command " alter database open resetlogs " from the target database to recover the database.
Instant recovery and database recovery
For general information on instant recovery, see the HPE Data Protector Concepts Guide and the HPE
Data Protector Zero Downtime Backup Administrator's Guide. For information on instant recovery in cluster environment (Cluster File System (CFS), HPE Serviceguard, and Microsoft Cluster Server), see the HPE Data Protector Zero Downtime Backup Administrator's Guide.
The Data Protector instant recovery functionality is used only to restore the target volumes on which the database files are located. The database recovery part is performed after instant recovery by the
RMAN utility. During database recovery, incremental backups and archive log backups performed after
ZDB to disk or ZDB to disk+tape are restored from tape. Only those archive logs that do not reside on the target volumes are restored.
IMPORTANT:
If the Oracle control file, online redo logs, and SPFILE are on the same source volumes as datafiles and you enable instant recovery by setting the omnirc
options, note that the control file, SPFILE, and online redo logs are overwritten during the instant recovery.
Data Protector (10.00) Page 102 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Prerequisites
l
The control file that reflects the internal database structure at the time of backup must be available on the application system. If necessary, restore the appropriate control file from a tape backup.
l
The recovery catalog must be open.
Limitations
l
For ZDB-to-disk sessions, only archived redo logs can be used for a database recovery after an instant recovery.
l
The recovery process will fail if the log entry with the specified logseq number or SCN number was created before the target volume.
RAC preparation steps
In case of RAC, set the following option in the omnirc
file:
ZDB_IR_VGCHANGE=vgchange -a s
The instant recovery procedure is the same as without RAC.
However, if instant recovery is to be performed to some other node than the one that was backed up, the following procedure must be performed before the standard instant recovery procedure:
1. Make sure that the HPE SG virtual package is running on the target node.
2. Set the OB2BARHOSTNAME environment variable as the virtual hostname before running the configuration from the command line: export OB2BARHOSTNAME=virtual_hostname
Instant recovery using the Data Protector GUI
To perform an instant recovery:
1. Put the Oracle database instance in a nomount state using sqlplus . In case of RAC, set all instances to nomount state.
For example: sqlplus sql> shutdown immediate sql> startup nomount sql> exit
2. In the Context List, click Instant Recovery.
3. Expand Oracle Server and select the ZDB-to-disk or ZDB-to-disk+tape session from which you want to perform the restore.
4. In the Source tab, select the objects to recover. Only whole databases can be selected. With
HPE P9000 XP Disk Array Family, it is recommended to leave the Keep the replica after the
restore option set to enable a restart of an instant recovery session. With HPE P6000 EVA Disk
Array Family, replica is kept on the disk array only if the Copy replica data to the source
location is selected.
Set other HPE P6000 EVA Disk Array Family or HPE P9000 XP Disk Array Family options. For details, press F1.
Data Protector (10.00) Page 103 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Selecting backup sessions (P9000 XP Array example)
5. At this point, you can decide whether to perform a database recovery immediately after an instant recovery or not: l
To perform only an instant recovery, click Restore.
NOTE:
You can perform a database recovery at a later time either from the Data Protector
.
l
To perform a database recovery immediately after an instant recovery, click on the Options tab, select Recovery and then select the database recovery options. For a recovery until a selected time, logseq/thread number, or SCN number, it is recommended to reset the log files.
See Oracle recovery options, below
and Restore, recovery, and duplicate options , on page 86
for details on available options.
Oracle recovery options
Data Protector (10.00) Page 104 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
6. Click Restore or Preview. Note that preview only checks if the replica can be restored. It does not check if the database recovery will be successful.
Data Protector recovers the database after performing instant recovery by switching the database to a mount state, restoring the necessary incremental backups and archived redo logs from tape, and applying the redo logs.
If you reset the logs, reset the database; otherwise, Oracle will during the next backup try to use the logs that were already reset and the backup will fail. Login to the target and recovery catalog database and execute: rman target Target_Database_Login catalog Recovery_Catalog_Login
RMAN> RESET DATABASE;
RMAN> exit
Oracle database recovery after the instant recovery
To recover the Oracle database after the instant recovery has been performed, perform the following steps:
1. Put the Oracle database in a mount state by connecting to the target database from the sqlplus and then running the following command: startup mount
2. To recover the database, the following two options are available:
Data Protector (10.00) Page 105 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Perform a recovery from the Data Protector Manager Restore Context:
a. Expand Oracle Server and select the database to recover. In the Source tab, under
Restore action, select Perform recovery only.
Selecting the database for recovery
.
c. Click Restore.
l
Perform a manual database recovery using RMAN.
Run the following RMAN script to recover the database: run {
allocate channel
dev1 type 'sbt_tape' parms
'SBT_LIBRARY=/opt/omni/lib/libob2oracle8.so,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)';
recover database;
sql 'alter database open';
release channel
dev1;
}
For additional examples on how to recover the database after an instant recovery, see
Restoring Oracle using RMAN , on page 90 .
Data Protector (10.00) Page 106 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Oracle in Veritas Cluster instant recovery
If Oracle on the application system runs in a Veritas Cluster, the following two Veritas Cluster resources must be disabled before instant recovery is performed, and enabled after instant recovery has finished to prevent the failover of the Oracle Veritas Cluster Service Group: l
Veritas Cluster application resource for the Oracle application and l
Veritas Cluster mountpoint resource for the Oracle database files.
Follow the steps below to perform an instant recovery to the application system with Oracle in a Veritas
Cluster:
1. On the application system, enter the following commands to disable the two Veritas Cluster resources:
a. hares -offline application_resource_name -sys system where
application_resource_name
is the name of the Veritas Cluster application resource for the Oracle application and
system
is the name of the active node.
hares -offline mountpoint_resource_name -sys system where
mountpoint_resource_name
is the name of the Veritas Cluster mountpoint resource for the Oracle database files and
system
is the name of the active node.
b. hares -modify application_resource_name Enabled 0 where
application_resource_name
is the name of the Veritas Cluster application resource for the Oracle application.
hares -modify mountpoint_resource_name Enabled 0 where
mountpoint_resource_name
is the name of the Veritas Cluster mountpoint resource for the Oracle database files.
2. Perform an instant recovery.
3. If you performed only an instant recovery without the database recovery, use RMAN as described
in Oracle database recovery after the instant recovery , on page 105 to bring the Oracle database
to a consistent state.
4. On the application system, enter the following commands to enable the two Veritas Cluster resources:
a. hares -modify mountpoint_resource_name Enabled 1 where
mountpoint_resource_name
is the name of the Veritas Cluster mountpoint resource for the Oracle database files.
hares -modify application_resource_name Enabled 1 where
application_resource_name
is the name of the Veritas Cluster application resource for the Oracle application.
b. hares -online application_resource_name -sys system where
application_resource_name
is the name of the Veritas Cluster application resource for the Oracle application and
system
is the name of the active node.
hares -online mountpoint_resource_name -sys system where
mountpoint_resource_name
is the name of the Veritas Cluster mountpoint resource for the Oracle database files and
system
is the name of the active node.
Data Protector (10.00) Page 107 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Aborting sessions
You can abort currently running sessions by clicking the abort button.
If, during a session, RMAN or SQL*Plus do not respond when requested, Data Protector automatically aborts the session. By default, Data Protector waits for the response for 5 minutes. Using omnirc options or environment variables OB2_RMAN_COMMAND_TIMEOUT and OB2_SQLP_SCRIPT_TIMEOUT , you can modify this time interval.
details of how to set the corresponding omnirc
options, see the HPE Data Protector Help index:
“omnirc option”. Note that environment variables override omnirc options.
Troubleshooting
This section contains a list of general checks and verifications and a list of problems you might
encounter when using the Data Protector Oracle integration. You can start at Problems, on page 114
and if you cannot find a solution there, perform general checks and verifications.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
For general ZDB, restore, and instant recovery troubleshooting information, see the troubleshooting sections in the HPE Data Protector Zero Downtime Backup Administrator's Guide.
Before you begin
l
Ensure that the latest official Data Protector patches are installed. See the HPE Data Protector Help index: “patches” on how to verify this.
l
See the HPE Data Protector Product Announcements, Software Notes, and References for general
Data Protector limitations, as well as recognized issues and workarounds.
l
See https://softwaresupport.hpe.com/manuals for an up-to-date list of supported versions, platforms, and other information.
Checks and verifications
For more detailed information on performing any of the following procedures, see the Oracle documentation.
If your configuration, backup, or restore failed: l
On the application system, verify that you can access the Oracle target database and that it is opened:
1. Perform the following:
Windows systems: Set the
ORACLE_HOME
and
DB_NAME
variables.
UNIX systems: Export the
ORACLE_HOME
and
DB_NAME
variables as follows:
Data Protector (10.00) Page 108 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration o
If you are using an sh -like shell, enter the following commands:
ORACLE_HOME="ORACLE_HOME" export ORACLE_HOME
DB_NAME="DB_NAME" export DB_NAME o
If you are using a csh -like shell, enter the following commands: setenv ORACLE_HOME "ORACLE_HOME" setenv DB_NAME "DB_NAME"
2. Start SQL*Plus from the bin directory in the
ORACLE_HOME
directory: sqlplus /nolog
3. Start SQL*Plus and type: connect user_name/password@service as sysdba; select * from dba_tablespaces; exit
If this fails, open the Oracle target database.
l
On the application system, verify that you can access the recovery catalog (if used) and that it is opened as follows:
1. Export or set the
ORACLE_HOME
and
DB_NAME
variables as described in Perform the following:, on the previous page
.
2. Start SQL*Plus from the bin directory in the
ORACLE_HOME
directory: sqlplus /nolog
3. Start SQL*Plus and type: connect Recovery_Catalog_Login select * from rcver; exit
If this fails, open the recovery catalog.
l
Verify that the listener is correctly configured for the Oracle target database and the recovery catalog database. This is required to properly establish network connections:
1. Export or set the
ORACLE_HOME
variable as described in Perform the following:, on the previous page .
2. Start the listener from the bin directory in the
ORACLE_HOME
directory: lsnrctl status service
If this fails, startup the listener process and see the Oracle documentation for instructions on how to create a configuration file ( LISTENER.ORA
).
On Windows, the listener process can be started in the Control Panel > Administrative
Tools > Services .
Data Protector (10.00) Page 109 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Checking the status of the Oracle listener
The status of the respective listener service in the Services window should be Started, otherwise you must start it manually.
3. Start SQL*Plus from the bin
directory in the
ORACLE_HOME
directory: sqlplus /nolog
4. Start SQL*Plus and type: connect Target_Database_Login exit and then connect Recovery_Catalog_Login exit
If this fails, see the Oracle documentation for instructions on how to create a configuration file
( NAMES.ORA
).
l
Verify that the Oracle target database configured to allow remote connections with the system privileges:
1. Export or set the
ORACLE_HOME
and
DB_NAME
variables as described in Perform the following:, on page 108 .
2. Start SQL*Plus from the bin directory in the
ORACLE_HOME
directory: sqlplus /nolog
3. Start SQL*Plus and type: connect Target_Database_Login as SYSDBA exit
Repeat the procedure using SYSOPER instead of SYSDBA .
If this fails, see the Oracle documentation for instructions on seting up the password file and any relevant parameters in the initDB_NAME.ora
file.
l
From the application system, verify that the target database and recovery catalog database are configured to allow remote connections with the system privileges and to allow backup:
Data Protector (10.00) Page 110 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration o
If you use the recovery catalog database:
Export or set the
ORACLE_HOME
and
DB_NAME
variables as described in Perform the following:, on page 108
.
SQL> connect login_to_recovery_catalog_or_target_database as sysdba;
> exit
Recovery_Catalog
l
If you use the recovery catalog database, verify that the target database is registered in the recovery catalog:
1. Export or set the
ORACLE_HOME
variable as described in Perform the following:, on page 108 .
2. Start SQL*Plus from the bin directory in the
ORACLE_HOME;
directory: sqlplus /nolog
3. Start SQL*Plus and type: connect Recovery_Catalog_Login; select * from rc_database; exit
If this fails, start the configuration using Data Protector on the application system, or see the Oracle documentation for information on how to register an Oracle target database in the recovery catalog database.
l
On the application system, verify backup and restore directly to disk using an RMAN channel type disk:
If you use the recovery catalog:
1. Export or set the
ORACLE_HOME
variable as described in Perform the following:, on page 108 .
2. Start RMAN from the bin directory in the
ORACLE_HOME
directory: rman target Target_Database_Login catalog Recovery_Catalog_Login
If you do not use the recovery catalog:
1. Export or set the
ORACLE_HOME
variable as described in Perform the following:, on page 108 .
2. Start RMAN from the bin directory in the
ORACLE_HOME
directory:
An example of the RMAN backup script is presented below: run { allocate channel 'dev0' type disk; backup tablespace tablespace_name format
'
ORACLE_HOME/tmp/datafile_name';
}
After a successful backup, try to restore the backed up tablespace by executing the following restore script: run { allocate channel 'dev0' type disk; sql 'alter tablespace
tablespace_name offline immediate';
Data Protector (10.00) Page 111 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration restore tablespace tablespace_name; recover tablespace tablespace_name; sql 'alter tablespace tablespace_name online';
}
If this fails, see the Oracle documentation for details of how to execute a backup and restore directly to disk using RMAN.
Additionally, if your configuration or backup failed: l
Verify that the Data Protector software has been installed properly.
For details, see the HPE Data Protector Installation Guide.
l
Check if the SYSDBA privilege is granted to the Oracle administrator.
l
If you have special Oracle environment settings, ensure that they are entered in the Data Protector
Oracle configuration files on the Cell Manager. For information on setting the variables in the Data
Protector Oracle configuration files, see the util_cmd man page or the HPE Data Protector
Command Line Interface Reference.
l
Perform a filesystem backup (non-ZDB) of the Oracle Server system so that you can eliminate any potential communication problems between the Oracle Server and the Data Protector Cell Manager system.
For details on performing filesystem backups, see the HPE Data Protector Help index: “standard backup procedure”.
Ensure that the hostname defined in the backup specification as a system to be backed up is the name of the application system.
l
Windows systems: Check the
Data Protector Inet
service startup parameters on the Oracle
Server system:
Go to Control Panel > Administrative Tools > Services > Data Protector Inet.
The service must run under a specified user account. Make sure that the same user is also added to the Data Protector admin or user group.
l
Examine the system errors reported in the following file on the application system (Oracle proxycopy ZDB method) or backup system (Oracle backup set ZDB method) into the debug.log
file.
Additionally, if your backup or restore failed: l
Test the Data Protector internal data transfer using the testbar2 utility:
1. Verify that the Cell Manager name is correctly defined on the Oracle Server system. Check the cell_server file located in the default Data Protector client configuration directory, which contains the name of the Cell Manager system.
2. From the bin directory in the
ORACLE_HOME
directory, execute:
If backup failed:
specification_name
If restore failed:
testbar2 -type:Oracle8 -appname:DB_NAME-perform:restore
The hostname should not be specified in the object option. It is automatically provided by testbar2 .
3. You should see only NORMAL
messages displayed on your screen, otherwise examine the errors
Data Protector (10.00) Page 112 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration reported by the testbar2 utility by clicking the Details button in the Data Protector Monitor context.
If the messages indicate problems on the Data Protector side of the integration, proceed as follows: o
Check if the user under which the backup or restore session was started has appropriate Oracle permissions (for example, belongs to the DBA group). This user must also be in the Data Protector operator or admin user group.
o
Check that the respective Data Protector user group has the See private objects user right enabled.
o
If backup failed: Create an Oracle backup specification to back up to a null device or file. If the backup succeeds, the problem may be related to the backup devices. See the HPE Data
Protector Troubleshooting Guide for instructions on troubleshooting devices.
o
If restore failed: Execute the omnidb
command to see objects in the database.
If the test fails again, call a support representative for assistance.
Additionally, if your restore failed: l
Verify that an object exists on the backup media.
This can be done by executing the following command on the Oracle server system from the bin directory in the
ORACLE_HOME;
directory: omnidb -oracle8 "object_name" -session "Session_ID" -media
The output of the command lists detailed information about the specified Oracle object, as well as the session IDs of the backup sessions containing this object and a list of the media used. For detailed syntax of the omnidb command, see its man page.
l
Ensure that the database is in the correct state.
If you are trying to restore a database item using the Data Protector GUI and the GUI stops responding, try one of the following: o
If you are restoring the control file, the database should be in the
NoMount
state.
Open a command window and enter the following: sqlplus/nolog
SQL>connect user/password@service as sysdba
SQL>shutdown immediate
SQL>startup nomount o
If you are restoring datafiles, the database should be in the Mount state.
Open a command window and enter the following: sqlplus/nolog
SQL>connect user/password@service as sysdba
SQL>shutdown immediate
SQL>startup mount l
If there is a problem you cannot resolve while you are trying to restore a database item using the
Data Protector (10.00) Page 113 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Data Protector GUI, try using the RMAN CLI to restore the database items.
For information, see Restoring Oracle using RMAN , on page 90 .
l
Try putting the database into the Open state manually after using the Data Protector GUI to recover and restore a backup session.
If you have used the Data Protector GUI to recover and restore a backup session and you see the following error message:
Oracle Error: ORA-1589: must use RESETLOGS or NORESETLOGS option for database open.
Open a SQLplus window and use the following command: sqlplus/nolog
SQL>connect user/password@service as sysdba
SQL>alter database open noresetlogs;
If this does not work, try using the following command:
SQL>alter database open resetlogs;
Problems
Problem
SQL*Plus is unable to connect to destination
Action
Check if the Oracle listener process is up and running. Check if there are any environment variables you need to enter (for example, TNS_ADMIN ). Enter these variables in the Data Protector Oracle configuration files on the Cell Manager. For information, see the util_cmd man page.
Problem
The following error is displayed: ORA-12532: : invalid argument
If this is reported by SQL*Plus in the Data Protector monitor, the application system may be low on resources (CPU, memory, and so on).
Action
Try to configure the application system in such a way that it consumes as little resources as possible.
This error can be reproduced without using Data Protector by starting SQL*Plus on the application system, and connecting to the target database on the application system.
Problem
Backup set ZDB is aborted after 10 minutes
While performing a backup set ZDB, the following warning is displayed for each database datafile:
RMAN-06554: WARNING: file n is in backup mode
The ZDB session then aborts with the following message:
Bar backup session was started but no client connected in 600 seconds.
Data Protector (10.00) Page 114 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Action
Increase the value of the following global options (by default, these options are set to 10): l
If you upgraded Data Protector from a previous version of Data Protector:
SmWaitForFirstClient=minutes l
If you performed a clean installation:
SmWaitForFirstBackupClient=minutes
For more information about the global options, see the HPE Data Protector Troubleshooting Guide.
Problem
Backup set ZDB fails after changing the physical schema of the Database
Backup fails after you have modified the physical schema of a database, for example, if you added or dropped a tablespace, added a new datafile, or added or dropped a rollback segment. Depending on the performed modification, different error messages are displayed, for example:
RMAN-06056: could not access datafile datafile
The problem occurs because the physical schema of target database is not updated in the recovery catalog.
Action
Manually re-synchronize the recovery catalog database with the current control file.
Problem
On UNIX systems, a backup set ZDB-to-disk+tape session fails
While performing a backup set ZDB to disk+tape, the session fails with the following error when Oracle
Server attempts to start up an instance on the backup system:
[Major] From: [email protected] DB_NAME Time: Date Time
The database reported error while performing requested operation.
This problem occurs when either the user ID or the group ID number of the Oracle operating system user account on the application and backup systems do not match. Under such circumstances, Oracle
Server is unable to start up the instance on the backup system due to missing privileges.
Action
, and restart the session.
Problem
Proxy copy restore fails
Proxy copy restore fails with the following error:
RMAN-10035: exception raised in RPC: ORA-27197: skgfprs: sbtpcrestore returned error
RMAN-10031: ORA-27197 occurred during call to DBMS_BACKUP_
RESTORE.PROXYRESTOREDATAFILE
Data Protector (10.00) Page 115 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Action
Check the IDB for the session and the objects of the latest backup. You might check if a more recent session exists in the recovery catalog. Connect to the RMAN prompt: rman target user/password@TGT_DB catalog user/password@CDB
At the RMAN> prompt, enter list backup; to display a list of the objects in the recovery catalog. Check the list of Proxy Copy sessions, listed at the end.
To synchronize the recovery catalog and the IDB, execute the RMAN command: resync catalog;
After the synchronization is performed, restore should be possible.
Problem
Restore after a switch between ZDB methods fails
If you perform a restore to a specified time ( T_RESTORE ) that lies in the time interval between the start of the first backup of the entire database using the new method ( T_START ), and before this backup is finished ( T_FINISH ), RMAN may try to restore the backup files made with the new method using a channel allocated for backup files made using the previous method. As a result, the restore procedure fails.
Restore after a switch between ZDB methods fails
Action
Restore the backup session manually using RMAN scripts. Add the required parameter to the allocated channels, that is
OB2PROXYCOPY=1
for the channel which will be used for restoring the backup made using the proxy-copy ZDB method. Then restore the backup files using the correct channels.
For example, if you switched from the backup set to the proxy-copy ZDB method, the script may look similar to the following one: run {
ALOCATE CHANNEL 'dev_0' TYPE 'sbt_tape'
PARMS 'SBT_LIBRARY=
Path_to_Data_Protector_MML,
Data Protector (10.00) Page 116 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
ENV(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME)';
ALOCATE CHANNEL 'dev_1' TYPE 'sbt_tape'
PARMS 'SBT_LIBRARY= Path_to_Data_Protector_MML,
ENV(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME,
OB2PROXYCOPY=1)';
RESTORE DATAFILE
list_of_backup_set_backups UNTIL
T_RESTORE CHANNEL 'dev_0';
RESTORE DATAFILE list_of_proxy-copy_backups UNTIL
T_RESTORE CHANNEL 'dev_1';
RELEASE 'dev_0';
RECOVER DATABASE UNTIL T_RECOVER ...
RELEASE 'dev_1';
}
Where:
T_RESTORE specifies the time to which to restore and T_RECOVER the time to which to apply the transactions.
method.
the entire database ( T_START ) and before T_RESTORE.
Problem
Data Protector reports errors when calling SYS.LT_EXPORT_PKG.schema_inf_exp during
Oracle backup
The following errors are listed in the Data Protector monitor:
EXP-00008: ORACLE error 6550 encountered
ORA-06550: line 1, column 13:
PLS-00201: identifier 'SYS.LT_EXPORT_PKG' must be declared
ORA-06550: line 1, column 7:
PL/SQL: Statement ignored
EXP-00083: The previous problem occurred when calling
SYS.LT_EXPORT_PKG.schema_info_exp
. exporting statistics
Export terminated successfully with warnings.
[Major] From: ob2rman.pl@machine "MAKI" Time: 10/01/01 16:07:53
Export of the Recovery Catalog Database failed.
Action
Start SQL*Plus and grant the execute permission to the LT_EXPORT_PKG as follows (make sure that the user sys
has the
SYSDBA
privilege granted beforehand): sqlplus 'sys/password@CDB as sysdba'
SQL> grant execute on sys.lt_export_pkg to public;
Restart the failed backup session.
Data Protector (10.00) Page 117 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Problem
On a UNIX system, Data Protector reports “Cannot allocate/attach shared memory”
Backup fails and the following error message is displayed:
Cannot allocate/attach shared memory (IPC Cannot Allocate Shared Memory Segment)
System error: [13] Permission denied) => aborting
Action
Set the OB2SHMEM_IPCGLOBAL omnirc option to 1 to use the memory windowing properly, and restart the
failed backup session. See the HPE Data Protector Troubleshooting Guide for details on using the omnirc
file.
Problem
Backup fails after a point in time restore and recovery
The following error is displayed:
RMAN-06004: ORACLE error from recovery catalog database: RMAN-20003: target database incarnation not found in recovery catalog
Action
Connect to the target and recovery catalog database using RMAN and reset the database to register the new incarnation of database in the recovery catalog: rman target Target_Database_Login catalog Recovery_Catalog_Login
RMAN> RESET DATABASE;
RMAN> exit
Problem
Oracle online backup fails with the following error:
RMAN-06004: ORACLE error from recovery catalog database: RMAN-20220: controlfile copy not found in the recovery catalog
When running an online backup, Data Protector adds the filename of the controlfilecopy to the RMAN backup script. This filename has to be cataloged to the RMAN catalog prior to the backup command.
Action
To catalog the controlfilecopy to the RMAN catalog:
1. Connect to RMAN on the application system.
2. Execute the following command:
RMAN> catalog controlfilecopy 'CONTROL_FILE_LOCATION/ctrlDB_NAME.ctl'
Problem
Backup of archive logs on RAC cannot be performed
Data Protector (10.00) Page 118 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
On RAC, the archive logs are not installed on a NFS mounted disk. Backup of archive logs cannot be performed.
Action
Edit the archive logs backup specification: l
Add an additional allocate channel command for each node. l
Add a command to connect to each instance. The connection parameters should be given as
username
/
passwd
@
INSTANCE
.
For example, if you are using two nodes, the backup specification might look as follows: run { allocate channel 'dev_0' type 'sbt_tape' parms
'SBT_LIBRARY= Path_to_Data_Protector_MML,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME= DB_NAME,OB2BARLIST=RAC_arch)'
connect
username/passwd@INSTANCE_1; allocate channel 'dev_2' type 'sbt_tape' parms
'SBT_LIBRARY=
Path_to_Data_Protector_MML,
ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=
DB_NAME,OB2BARLIST=RAC_arch)' connect username/passwd@INSTANCE_2; backup
format 'RAC_arch<QU_%s:%t:%p>.dbf'
archivelog all;
}
Problem
The Recovery Catalog was lost and the control file cannot be restored from Data Protector managed backup
The Recovery Catalog was not used, the RMAN autobackup feature was not used, and the control file cannot be restored from Data Protector managed backup. A valid control file backup exists on tape.
Action
Restore the control file from RMAN backup set, mount and restore the database, and perform database recovery: run { allocate channel 'dev_0' type 'sbt_tape' parms
'SBT_LIBRARY=
Path_to_Data_Protector_MML'; restore controlfile from '
backup piece handle'; sql 'alter database mount'; set until time ' MMM DD YY HH24:MM:SS'; restore database; recover database; sql 'alter database open resetlogs'; release channel 'dev_0';
}
At this point you must manually register any backups made after the control file backup that was restored. After that, continue with the restore procedure.
Data Protector (10.00) Page 119 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
For the backup piece handle search the Data Protector Internal Database and session outputs of previous backup sessions.
Problem
How to modify the RMAN restore script
When you start a restore of an Oracle database using the Data Protector GUI or CLI, an RMAN restore script is created, which is instantly run, so you cannot edit it first.
Action
To edit the script before it is run, set the Data Protector omnirc option OB2RMANSAVE to point to an existing directory. When the variable is set and you start a restore, the RMAN restore script, which is created at run time, is saved to the specified location under the name RMAN_restore_backup_
specification_name.rman
, and the actual restore is skipped. Then you can edit the script and run it manually afterwards. On how to set the omnirc
options, see the HPE Data Protector Help index:
“omnirc options”.
To start a restore using Data Protector again, clear the OB2RMANSAVE option by deleting its content or commenting or removing the whole option. If you comment or remove the option on a Windows system, restart the Data Protector Inet service for the settings to take effect.
Problem
Instant recovery session for an Oracle database fails
On a Windows Server 2008 x64 system, when you run an instant recovery from a backed up Oracle
11g application database that resides on a disk array of the HPE P6000 EVA Disk Array Family, the session may fail with the following error messages:
[Major] From: [email protected] "SMISA"
Time: 5/17/2010 5:00:50 AM
A filesystem could not be mounted.
Filesystem name :
Mount point : I:\
[Critical] From: [email protected] "SMISA"
Time: 5/17/2010 5:00:50 AM
Failed to resume the application system.
[Critical] From: [email protected] "SMISA"
Time: 5/17/2010 5:00:50 AM
Instant Recovery failed.
Such an instant recovery failure occurs when the “copy-back” instant recovery method (the default) is chosen and the following instant recovery options are selected in the Source and Options panes of the
Data Protector GUI: l
Wait for the replica to complete (with the default waiting period used) l
Retain source for forensics l
Check the data configuration consistency l
Force the removal of all replica presentations
Data Protector (10.00) Page 120 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration l
Recovery l
Open database after recovery
Action
On the application system and the backup system, set the values of the omnirc options ZDB_DELAY_
BEFORE_RESCAN and ZDB_DELAY_AFTER_RESCAN to 300 and restart the instant recovery session.
Problem
Instant recovery of an Oracle database fails
An instant recovery session for an Oracle database fails with a message similar to the following:
[Normal] From: [email protected] "testdb" Time:
2/7/2008 10:48:19 AM
Starting target database instant recovery.
Net service name: testdb.
Instance status: .
Instance name: .
Database DBID = .
Database control file type: .
Database log mode: .
[Major] From: [email protected] "testdb" Time:
2/7/2008 10:48:20 AM
The database reported error while performing requested operation.
Note that the database parameters are empty, which happens when Data Protector is not able to connect to the Oracle database.
Action
Data Protector must be able to connect to the Oracle database even when the database is in the Mount or
Nomount
state. One way of achieving this is by configuring static service information for your Oracle listener. For details, see the Oracle documentation.
The following example shows how static service information is configured using the Oracle Net
Manager.
Suppose you have the following environment:
Listener name: LISTENER
Global database name: orcl
Oracle SID: ORCL
To configure static service information for the listener, open the Oracle Net Manager, select the listener, go to the Database Services context, add a database and specify the Oracle database parameters.
Data Protector (10.00) Page 121 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
Configuring an Oracle listener
As a result, the listener.ora
file gets updated with the SID_LIST_LISTENER section.
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(GLOBAL_DBNAME = orcl)
(SID_NAME = ORCL)
)
)
At the end, restart the Oracle listener service to apply the changes.
Problem
“IPC Invalid Hostname or IP Address” error message is displayed when browsing Oracle restore sessions
The following error message is displayed when browsing the Oracle database for restore sessions in the Data Protector GUI Restore context:
IPC Invalid Hostname or IP Address
The problem can appear in the following cases: l
When restoring database items to a different client.
l
When importing Data Protector media containing backups of Oracle database from another Data
Protector cell.
l
When restoring 64-bit Oracle version 10.2.0.4 in the RAC environment, on HP-UX 11.23 PA-RISC systems. If util_orarest is present on the system, this error can mean that the util_orarest agent ends abnormally while trying to load the 32-bit OCI library from the ORACLE_HOME/lib32 directory.
Actions
l
To successfully restore database items to a different client, make sure that the system on which the
Data Protector Oracle integration agent will be started is configured as Data Protector Oracle database instance (ORACLE_SID).
To verify, check if it is listed in the Client drop-down list in the Options page.
Data Protector (10.00) Page 122 of 413
Zero Downtime Backup Integration Guide
Chapter 1: Data Protector Oracle Server ZDB integration
of the Restoring Oracle database objects , on page 80 procedure.
l
When restoring 64-bit Oracle database in RAC environment, on HP-UX 11.23, resolve the problem as follows:
In the directory ORACLE_HOME/lib remove the soft link libclntsh.sl
, which points to the 64-bit
OCI library
ORACLE_HOME/lib/liblntsh.sl.10.1
.
Problem
When editing the RMAN script section of a backup specification using the Data Protector GUI, an RMAN backup script error is displayed
In the Data Protector GUI, when you edit the RMAN script section of a Data Protector backup specification, the following error message may display:
Cannot proceed, invalid RMAN backup script.
The error is displayed if you have specified an Oracle RMAN parameter, which has not been recognized by the Data Protector parser or a parsing error has occurred.
Action
Disable Oracle RMAN script parsing in the Data Protector GUI by setting the Data Protector
NoGUIRMANScriptParsing global option to 1 .
For details of how to set the option, see the HPE Data Protector Help index: "global options".
Problem
The Oracle Recovery Manager fails with "fatal error in recovery manager"
The Oracle RMAN fails with the following error messages:
RMAN-06900: WARNING: unable to generate V$RMAN_STATUS or V$RMAN_OUTPUT row
RMAN-06901: WARNING: disabling update of the V$RMAN_STATUS and V$RMAN_OUTPUT rows
RMAN-06003: ORACLE error from target database:
ORA-03113: end-of-file on communication channel
Process ID: 4222
Session ID: 29 Serial number: 51
..
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-00601: fatal error in recovery manager
RMAN-03004: fatal error during execution of command
Action
In the sqlnet.ora
file, increase the SQLNET.RECV_TIMEOUT
value.
Data Protector (10.00) Page 123 of 413
Chapter 2: Data Protector SAP R/3 ZDB integration
Introduction
This chapter explains how to configure and use the Data Protector SAP R/3 ZDB integration (SAP R/3 ZDB
integration). It describes concepts and methods you need to understand to back up and restore the following files of the SAP R/3 database environment (SAP R/3 objects): l
Data files l
Control files l
Online redo logs l
Offline (archived) redo logs l
SAP R/3 logs and parameter files
Data Protector supports offline and online backups. During an online backup, the SAP R/3 application is actively used.
Data Protector offers interactive and scheduled backups of the following types: l
ZDB to disk l
ZDB to tape l
ZDB to disk+tape
Data Protector supports only a filesystem restore. You can restore SAP R/3 files: l
To the original location l
To another client l
To another directory
Backup and restore sessions, below
shows which restore methods are available, depending on the ZDB session you restore from.
Backup and restore sessions
ZDB type Restore methods
ZDB to tape
ZDB to disk
ZDB to disk+tape
Standard restore
Instant recovery
Standard restore, instant recovery
When the instant recovery completes, you can recover the database to a specific point in time using the SAP
BRTOOLS interface.
Supported disk arrays and array configurations
Data Protector (10.00) Page 124 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Supported array
EMC Symmetrix (EMC)
Supported configurations
TimeFinder, SRDF, combined SRDF+TimeFinder
HPE P9000 XP Disk Array Family
(P9000 XP Array)
HPE P6000 EVA Disk Array Family
(P6000 EVA Array)
HPE BC P9000 XP, HPE CA P9000 XP, combined HPE
CA+BC P9000 XP
HPE BC P6000 EVA, combined HPE CA+BC P6000
EVA
Local replication Non–HPE Storage Arrays (NetApp storage)
This chapter provides information specific to the Data Protector SAP R/3 ZDB integration. For general
Data Protector procedures and options, see the HPE Data Protector Help. For details on ZDB terminology, ZDB types, advantages of offline and online backups, and instant recovery concepts, see the HPE Data Protector Concepts Guide.
Integration concepts
SAP R/3 integration architecture, below shows the architecture of the Data Protector SAP R/3 ZDB
integration. The figure illustrates the preferred configuration, in which the Oracle control file, online redo log files, and Oracle SPFILE reside on a different volume group than the Oracle data files.
SAP R/3 integration architecture
Data Protector (10.00) Page 125 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
For other supported configurations, see ZDB integrations omnirc options, on page 408
.
ZDB flow
For details of how ZDB options affect the ZDB flow, including mounting, activating volume or disk groups and so on, see the HPE Data Protector Zero Downtime Backup Administrator's Guide.
SAP R/3 ZDB session flow (BRBACKUP started on the application system)
1. The SAP R/3 backup specification is read and the Data Protector omnisap.exe
program is started on the application system.
2. omnisap.exe prepares an exclude list (Oracle control files and online redo log files)
3. Omnisap.exe
starts BRBACKUP, which switches the database to the backup mode (online backup) or shuts down the database (offline backup), and starts the split command on the application system, with the list of files to be included in the replica creation.
The database is put out of the backup mode (online backup) or is restarted (offline backup) after the replica is created.
NOTE:
Data Protector can use the splitint interface (if BRTOOLS supports it) to reduce the time during which the database is in the backup mode.
4. The Data Protector ob2smbsplit command resolves the backup configuration, creates a replica,
Data Protector (10.00) Page 126 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration and preparing the replica for backup.
The mountpoints for the backed up object are created on the backup system.
The backup volume/disk groups are activated and the filesystems are mounted on the backup system.
5. ZDB to disk: The remaining ZDB options are processed and the details on the session are written to the ZDB database. The session then finishes.
6. ZDB to tape, ZDB to disk+tape: BRBACKUP starts the Data Protector backint
program, which starts establishing a connection between the Data Protector Data Movement Agents (DMA) on the backup system and the General Media Agents (MA). The process is coordinated by the
Data Protector Backup Session Manager (BSM). When the connection is established, the data specified for backup is streamed to tape.
NOTE:
You can configure SAP to use a third-party backint tool to perform ZDB-to-disk+tape backup sessions. To enable this feature, set the
OB2_3RD_PARTY_BACKINT
environment variable to
1
and copy the backint
you want to use to the SAP BRTOOLS directory. In case of a ZDB-to-disk+tape backup session, a disk_only backup object is created in the
Data Protector IDB. The third-party backint starts after the split and is responsible for backing up the needed files.
7. When the data transfer completes, the backup system is disabled (filesystems are dismounted on all platforms and volume/disk groups deactivated on UNIX).
8. EMC and P9000 XP Array: Links are re-established, depending on how ZDB options are specified.
Data Protector SAP R/3 configuration file
Data Protector stores the integration parameters for every configured SAP R/3 database in the following file on the Cell Manager:
Windows systems:
Data_Protector_program_
data\Config\Server\Integ\Config\Sap\ClientName%ORACLE_SID
UNIX systems: /etc/opt/omni/server/integ/config/SAP/ClientName%ORACLE_SID
The parameters stored are: l
Oracle home directory l encoded connection string to the target database l
BRTOOLS home directory l the variables which need to be exported prior to starting a backup l
SAPDATA home directory l user name and user group l temporary directory used for the copy of the control file or redo logs l list of control files and redo logs that will be copied to a safe location
Data Protector (10.00) Page 127 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration l concurrency number and balancing (for each backup specification), and number of channels for
RMAN backup l speed parameters (time needed for a specific file to back up - in seconds) l manual balancing parameters
The configuration parameters are written to the Data Protector SAP R/3 configuration file: l during configuration of the integration l during creation of a backup specification l when the configuration parameters are changed
IMPORTANT:
To avoid problems with your backups, take extra care to ensure the syntax and punctuation of your configuration file match the examples.
NOTE:
You can set up the parameters in the Environment section (sublist) of the file by referring to other environment variables in the following way:
SAPDATA_HOME=${ORACLE_HOME}/data
Syntax
The syntax of the Data Protector SAP R/3 configuration file is as follows:
ORACLE_HOME='
ORACLE_HOME';
ConnStr=' ENCODED_CONNECTION_STRING_TO_THE_TARGET_DATABASE';
BR_directory=' BRTOOLS_HOME';
SAPDATA_HOME=' SAPDATA_HOME';
ORA_NLS_CHARACTERSET=' CHARACTER_SET';
OSUSER='
USER_NAME';
OSGROUP='
USER_GROUP';
Environment={
[
ENV_var1='value1';]
[ ENV_var2='value2';
...]
}
SAP_Parameters={ backup_spec_name=('-concurrency #_of_concurrency
' | '-time_balance' | '-load_balance' | '-manual_balance');
} speed={
AVERAGE=1;
' filename'=#_of_seconds_needed_to_back_up_this_file;
} compression={' filename'=size_of_the_file_in_bytes_after_the
_compression;
} manual_balance={
backup_specification_name={
'
filename'=device_number;
}
}
Data Protector (10.00) Page 128 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
The
ORA_NLS_CHARACTERSET
parameter is set automatically by Data Protector during SAP R/3 database configuration. For details of how to configure SAP R/3 database for use with Data Protector,
see Configuring SAP R/3 databases, on page 139 .
Example
This is an example of the file:
ORACLE_HOME='/app/oracle805/product';
ConnStr='EIBBKIBBEIBBFIBBGHBBOHBB
QDBBOFBBCFBBPFBBCFBBIFBBGFBBDGBBBFBBCFBBDFBBCFBB';
BR_directory='/usr/sap/ABA/SYS/exe/run';
SAPDATA_HOME='/sap';
ORA_NLS_CHARACTERSET='USASCII7';
OSUSER='orasid';
OSGROUP='dba';
Environment={ }
SAP_Parameters={
sap_weekly_offline=('-concurrency 1','-no_balance');
sap_daily_online=('-concurrency 3','-load_balance');
sap_daily_manual=('-concurrency 3','-manual_balance');
}
speed={
AVERAGE=203971;
'/file1'=138186;
'/file2'=269756;
}
compression={
'/file1'=1234;
'/file2'=5678;
}
manual_balance={
sap_daily_manual={
'/file1'=1; /* file 1 is backed up by the first sapback */
'/file2'=2; /* file 2 is backed up by the second sapback */
'/file3'=1; /* file 3 is backed up by the first sapback */
'/file4'=1;
}
}
Setting, retrieving, listing, and deleting Data Protector
SAP R/3 configuration file parameters using the CLI
The Data Protector SAP R/3 configuration file parameters are normally written to the Data Protector
SAP R/3 configuration file after: l the Data Protector configuration of the Oracle instance that is run by SAP R/3 is completed.
l a new backup specification is created.
Data Protector (10.00) Page 129 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration l a backup that uses balancing by time algorithm is completed.
The
util_cmd
command
You can set, retrieve, list, or delete the Data Protector SAP R/3 configuration file parameters using the util_cmd -putopt
(setting a parameter), util_cmd -getopt
(retrieving a parameter), or util_cmd getconf
(listing all parameters) command on the Data Protector SAP R/3 client.
Cluster-aware clients
In a cluster environment, the environment variable
OB2BARHOSTNAME
must be defined as the virtual hostname before executing the util_cmd command from the command line (on the client). The
OB2BARHOSTNAME variable is set as follows:
Windows systems: set OB2BARHOSTNAME=virtual_hostname
UNIX systems: export OB2BARHOSTNAME=virtual_hostname
The
util_cmd
synopsis
The syntax of the util_cmd
command is as follows: util_cmd -getconf[ig] SAP oracle_instance [-local filename] util_cmd -getopt[ion] [SAP oracle_instance] option_name [-sub[list] sublist_name]
[-local filename] util_cmd -putopt[ion] [SAP oracle_instance] option_name [option_value] [-sub[list]
sublist_name] [-local filename] where:
[-sub[list] sublist_name]
specifies the sublist in the configuration file to which a parameter is written to or taken from.
[-local filename] specifies one of the following: l
When it is used with the -getconf[ig] option, it specifies the filename for the output of the command to be written to. If the -local option is not specified, the output is written to the standard output.
l
When it is used with the -getopt[ion] , it specifies the filename of the file from which the parameter
and its value are to be taken and then written to the standard output. If the -local option is not specified, the parameter and its value are taken from the Data Protector SAP R/3 configuration file and then written to the standard output.
l
When it is used with the -putopt[ion] option, it specifies the filename for the output of the command to be written to. If the -local option is not specified, the output is written to the Data
Protector SAP R/3 configuration file.
NOTE:
If you are setting the
option_value
parameter as a number, the number must be put in single quotes, surrounded by double quotes.
Data Protector (10.00) Page 130 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Return values
The util_cmd
command displays a short status message after each operation (writes it to the standard
error): l
Configuration read/write operation successful.
This message is displayed when all the requested operations have been completed successfully.
l
Configuration option/file not found.
This message is displayed when either an option with the specified name does not exist in the configuration, or the file specified as the -local parameter does not exist.
l
Configuration read/write operation failed.
This message is displayed if any fatal errors occurred, for example: the Cell Manager is unavailable, the Data Protector SAP R/3 configuration file is missing on the Cell Manager, and so on.
Setting parameters
To set the Data Protector OB2OPTS and the Oracle BR_TRACE parameters for the Oracle instance ICE that is run by SAP R/3, use the following commands on the Data Protector SAP R/3 client:
Windows, HP-UX, Solaris systems
util_cmd -putopt SAP ICE OB2OPTS '-debug 1-200 debug.txt' -sublist Environment util_cmd -putopt SAP ICE BR_TRACE "'10'" -sublist Environment
Retrieving parameters
To retrieve the value of the
OB2OPTS
parameter for the Oracle instance
ICE
, use the following command on the Data Protector SAP R/3 client: util_cmd -getopt SAP ICE OB2OPTS -sublist Environment
Listing parameters
To list all the Data Protector SAP R/3 configuration file parameters for the Oracle instance ICE , use the following command on the Data Protector SAP R/3 client: util_cmd -getconf SAP ICE
Deleting parameters
To remove the value of the OB2OPTS parameter for the Oracle instance ICE , use the following command on the Data Protector SAP R/3 client: util_cmd -putopt SAP ICE OB2OPTS "" —sublist Environment
Configuring the integration
To configure the integration:
1. Configure the required user accounts. See Configuring user accounts, on page 133 .
Data Protector (10.00) Page 131 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
.
4. Enable the use of the authentication password file. See Authentication password file, on page 136
.
.
7. Configure every SAP R/3 database you intend to back up from or restore to. See Configuring SAP
8. Configure the SAP R/3 parameter file. See Configuring the SAP R/3 parameter file , on page 146
.
Prerequisites
l
Ensure that you have correctly installed and configured the SAP R/3 application. The database used by the SAP R/3 application must be an Oracle database. If any other database is used, you can back it up using the corresponding Data Protector integration. It is assumed that you are familiar with the SAP R/3 application and Oracle database administration.
o
For supported versions, platforms, devices, and other information, see the latest support matrices at https://softwaresupport.hpe.com/manuals .
o
For information on installing, configuring, and using the SAP R/3 application and the SAP backup and restore tools (BRBACKUP, BRRESTORE, and BRARCHIVE), see the SAP R/3 application documentation.
l
Ensure that you have a license to use the Data Protector SAP R/3 ZDB integration. For information, see the HPE Data Protector Installation Guide.
l
Ensure that you have correctly installed Data Protector.
o
For information on how to install the Data Protector disk array integration (P6000 EVA Array,
P9000 XP Array, EMC, or non-HPE Storage Arrays) with SAP R/3 in various architectures, see the HPE Data Protector Installation Guide.
o
On how to configure the Data Protector ZDB integration (EMC, P9000 XP Array, P6000 EVA
Array, or NetApp Storage), see the HPE Data Protector Zero Downtime Backup Administrator's
Guide.
o
For information on the Data Protector Cell Manager package configuration in the HPE SG cluster, see the HPE Data Protector Help index: “HPE Serviceguard integration”.
NOTE:
You cannot run ZDB sessions in the RMAN mode.
l
The SAP R/3 directories SAPBACKUP , SAPARCH , SAPREORG , SAPCHECK , and SAPTRACE must not reside on the same disk array source volumes as the data files. Otherwise, the BRTOOLS data needed for complete recovery of a database is overwritten during instant recovery. You can set the locations for these directories in the initDBSID.sap
file.
Data Protector (10.00) Page 132 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Before you begin
l
Configure devices and media for use with Data Protector.
l
To test whether the SAP R/3 system and the Cell Manager communicate properly, configure and run a Data Protector filesystem backup and restore.
l
Windows systems:
On supported Windows operating systems, configure the Data Protector Inet service user impersonation for the user that has the appropriate SAP R/3 permissions for running backups and restores.
For details, see the HPE Data Protector Help index: “Inet user impersonation”.
If there are several SAP R/3 instances running on the same system with different SAP administrator accounts configured for each instance, create an additional, common SAP administrator account.
Configure the Data Protector Inet service to use this account as the service startup account.
Cluster-aware clients
l
Configure SAP R/3 databases only on one cluster node, since the configuration files reside on the
Cell Manager.
Windows systems: During the configuration, Data Protector copies the Data Protector backint and ob2smbsplit.exe
programs (the latter only if splitint is supported by BRTOOLS) from
Data_
Protector_home\bin to the directory that stores the SAP backup tools and renames ob2smbsplit.exe
to splitint.exe
. This is done only on the currently active node. On the other node, do it manually.
UNIX systems: During the configuration, Data Protector creates a link to the Data Protector backint and splitint programs on the currently active node. On all the other nodes, do it manually.
Execute: ln -s /opt/omni/lbin/backint \ /usr/sap/ORACLE_SID/sys/exe/run
If splitint is supported by BRTOOLS, also execute: ln -s /opt/omni/lbin/ob2smbsplit \ /usr/sap/ORACLE_SID/sys/exe/run/splitint l
If you intend to use the Data Protector CLI, set the Data Protector environment variable
OB2BARHOSTNAME
to the virtual server name as follows:
Windows systems: set OB2BARHOSTNAME=virtual_server_name
UNIX systems: export OB2BARHOSTNAME=virtual_server_name
NOTE:
SAP recommends installing SAP backup utilities on all cluster nodes.
Configuring user accounts
To enable backup and restore of SAP R/3 database files, you need to configure or create several user accounts.
Data Protector (10.00) Page 133 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Oracle operating system user account
User account root
(UNIX systems only)
Oracle database user account
Operating system user account that is added to the following user groups:
Windows systems:
ORA_DBA
and
ORA_SID_DBA
local groups
UNIX systems: dba and sapsys
For example, user oraSID .
UNIX systems: Ensure that this user is the owner of the filesystem or of the raw logical volume on which the database is mounted. The minimum permissions should be 740.
Default operating system administrator's user account added to the dba user group.
Database user account granted at least the following Oracle roles: l sysdba l sysoper
For example, user system .
Do not configure the Oracle SYS user for backing up SAP R/3 objects.
When backing up using the SYS user account, the SAP backup fails with the error
ORA-28009: connection as SYS should be as SYSDBA or
SYSOPER.
Add the following user accounts to the Data Protector admin or operator user group: l
Oracle operating system user account
(if you are using backup set method, add this user on the application as well as on backup system) l
UNIX systems: User account root (for both the application system and backup system)
In cluster environments, add these user accounts to the Data Protector admin or operator user group for the following clients: l virtual server l every node in the cluster
For information on adding Data Protector users, see the HPE Data Protector Help index: “adding users”.
Configuring SQL*Net V2 or Net8 TNS listener
1. Ensure that the listener.ora
and tnsnames.ora
files on the application system are configured as shown in the following example. The files are located in:
UNIX systems: ORACLE_HOME/network/admin
Windows systems:
ORACLE_HOME\network\admin
Example
Oracle instance:
PRO
Application system: alpha.hp.com
Data Protector (10.00) Page 134 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration listener.ora
tnsnames.ora
LISTENER =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS =
(PROTOCOL = TCP) (HOST = alpha.hp.com)
(PORT = 1522)
)
)
)
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(GLOBAL_DBNAME = PRO)
(SID_NAME = PRO)
(ORACLE_HOME = /app/oracle815/product)
)
)
PRO =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)
(HOST = alpha.hp.com) (PORT = 1522))
)
(CONNECT_DATA = (SERVICE_NAME = PRO)
)
)
2. Start the SQL*Net V2 or Net8 TNS listener by executing the following on the Oracle Server system:
UNIX systems: ORACLE_HOME/bin/lsnrctl start
Windows systems: ORACLE_HOME\bin\lsnrctl start
Checking the connection
To check the connection to the Oracle instance from the application system:
1. Log in to the application system as the Oracle OS user.
2. Export/set the
ORACLE_HOME
and
ORACLE_SID
variables.
3. Start sqlplus .
4. Connect to the Oracle target database as the Oracle database user, first with the sysdba role and then with the sysoper role.
Example
For the following configuration:
Oracle instance: PROORACLE_HOME : /app/oracle816/product
Data Protector (10.00) Page 135 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration execute: id uid=102(oraprod) gid=101(dba) export ORACLE_SID=PRO export ORACLE_HOME=/app/oracle816/product export SHLIB_PATH=/app/oracle816/product/lib:/opt/omni/lbin sqlplus /nolog
SQLPLUS> connect system/manager@PRO as sysdba;
Connected.
SQLPLUS> connect system/manager@PRO as sysoper;
Connected.
Authentication password file
Enable the use of the authentication password file for the database administrator:
1. Shut down the Oracle target database on the application system.
2. In the init
ORACLE_SID
.ora
file, specify: remote_login_passwordfile = exclusive
For instructions on how to set up the password file, see the Oracle documentation.
Enabling archived logging
When you set the database to the archived logging mode, you protect the unsaved online redo logs from being overwritten. Online backup of data files is useless without the related redo logs because you cannot recover the database to a consistent state.
TIP:
Archive the redo log files generated during the online backup immediately after BRBACKUP completes.
To protect the archive directory from overflowing, clear the directory regularly.
To enable archived logging:
1. In the initORACLE_SID.ora
file, set log_archive_start = true and specify the log_archive_dest option.
Example
This is an example of the initORACLE_SID.ora
file for the Oracle instance PRO :
# @(#)initSID.ora 20.4.6.1 SAP 13/03/30
#####################################################
# (c)Copyright SAP AG, Walldorf
#####################################################
. . . .
. . . . . . . . . .
Data Protector (10.00) Page 136 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
. . . . . . . .
. . . . . . . .
### ORACLE Authentication Password File remote_login_passwordfile = exclusive
### ORACLE archiving log_archive_dest = /oracle/PRO/saparch/PROarch log_archive_start = true
. . . .
2. Mount the Oracle database and start the archived logging mode using the Oracle Server Manager.
Execute: startup mount alter database archivelog; archive log start; alter database open;
Example
For the Oracle instance PRO , execute:
Windows systems: set ORACLE_SID=PRO
UNIX systems: export ORACLE_SID=PRO
Any operating system:
sqlplus /nolog
SQLPLUS> connect user/passwd@PRO;
Connected.
SQLPLUS> startup mount
ORACLE instance started.
Total System Global Area 6060224 bytes
Fixed Size 47296 bytes
Variable Size 4292608 bytes
Database Buffers 1638400 bytes
Redo Buffers 81920 bytes
Database mounted.
SQLPLUS> alter database archivelog;
Statement processed.
SQLPLUS> archive log start;
Statement processed.
SQLPLUS> alter database open;
Sharing directories on the application system
The following directories on the application system must be accessible: l sapbackup l sapreorg l
Oracle home directory l
BR*Tools home directory
Data Protector (10.00) Page 137 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
NOTE:
The sapreorg and BR*Tools home directories must be accessible only if you want to run SAP compliant ZDB sessions (BRBACKUP is started on the backup system and not on the application system).
UNIX application system
1. Share the directories on the application system through NFS with root permissions.
For example, suppose that the sapbackup directory on the application system points to
/oracle/SID/sapbackup and the backup system is backup.company.com
. To share the sapbackup directory:
HP-UX systems: In the file /etc/exports on the application system, add the line:
/oracle/SID/sapbackup -root=backup.company.com
Solaris systems: In the file
/etc/dfs/dfstab
on the application system, add the line: share -F nfs -o root=backup.company.com/oracle/SID/sapbackup
2. Mount the directories on the backup system. Ensure that you have the same directory structure on both the application and backup system.
For example, suppose that you have an HP-UX application system app.company.com
. To mount the /oracle/SID/sapbackup directory on the backup system, add the following line to the file
/etc/fstab on the backup system: app.company.com:/oracle/
SID/sapbackup
/oracle/
SID/sapbackup nfs defaults 0 0
Windows application system
l
On the application system, find the location of the sapbackup
and sapreorg
directories. If they reside inside the SAP data home directory, share this directory under any name you want. Then, on the backup system, create the HKEY_LOCAL_MACHINE\SOFTWARE\SAP\ ORACLE_
SID\Environment\SAPDATA_HOME Windows registry key, specifying the SAP data home directory path as seen from the backup system.
For example, suppose your application system is mycomputer.company.com
and the SAP data home is K:\oracle\my_instance , which you share under the name my_SAPinstance . Then, the
SAPDATA_HOME Windows registry key on the backup system must have the value
\\mycomputer.company.com\my_SAPinstance .
If the sapbackup and sapreorg directories do not reside together, share each directory separately and also create a separate Windows registry key on the backup system ( SAPBACKUP , SAPREORG ).
l
Share the Oracle home directory on the application system and specify its path in the ORACLE_HOME
Windows registry key on the backup system, similarly as described above.
l
Ensure that the BR*Tools home directory on the application system is accessible from the backup system as follows:
\\application_system\sapmnt\SAP_SID\SYS\exe\run
TIP:
Data Protector (10.00) Page 138 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Instead of creating registry keys, you can also set the Data Protector SAP R/3 integration environment variables ( ORACLE_HOME , SAPDATA_HOME , SAPBACKUP , SAPREORG ).
Choosing authentication mode
Data Protector SAP R/3 ZDB integration supports two authentication modes for accessing Oracle databases that are used by SAP R/3: l database authentication mode l operating system authentication mode
With database authentication mode, you need to re-configure the SAP R/3 integration for an SAP R/3 database with the new Oracle login information each time the corresponding Oracle database user account changes. Such a reconfiguration is not needed if operating system authentication mode is used.
You select the preferred authentication mode when you configure a particular SAP R/3 database.
Configuring SAP R/3 databases
You need to provide Data Protector with the following configuration parameters: l
Oracle Server home directory l
SAP R/3 data home directory l
Optionally, if you choose database authentication mode, Oracle database user account. The user account is used by BRBACKUP and BRARCHIVE during backup.
l
Directory in which the SAP backup utilities are stored
Data Protector then creates the configuration file for the SAP R/3 database on the Cell Manager and verifies the connection to the database. On UNIX systems, Data Protector also creates a soft link for the backint program from the directory that stores the SAP backup utilities to /opt/omni/lbin .
On Windows systems, Data Protector copies the backint and programs (the latter only if splitint is supported by BR*Tools) from Data_Protector_home\bin to the directory that stores the SAP backup tools and renames ob2smbsplit.exe
to splitint.exe
.
To configure an SAP R/3 database, use the Data Protector GUI or CLI.
Before you begin
l
Ensure that the SAP R/3 database is open.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click SAP R/3, and click Add
Backup.
3. In the Create New Backup dialog box, select the template.
Data Protector (10.00) Page 139 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
From the Backup type drop-down list, select the Snapshot or split mirror backup option, and from the Sub type drop-down list, select the appropriate disk array agent. The agent must be installed on the application system and the backup system.
Click OK.
4. In Application system, select the SAP R/3 client to be backed up. In cluster environments, select the virtual server.
In Backup system, select the backup system.
Specify other disk array specific backup options (see EMC backup options, on page 150
for EMC,
P9000 XP Array backup options, on page 150
for P9000 XP Array, P6000 EVA Array backup options, on page 151
for P6000 EVA Array, or NetApp Storage backup options, on page 152 for
NetApp Storage). For details on the options, press F1.
NOTE:
P9000 XP Array and P6000 EVA Array: To enable instant recovery, select Track the
replica for instant recovery.
5. In Application database, type the Oracle instance name ( ORACLE_SID
).
Specify the User and group/domain options, which are available on UNIX and Windows Server
2008 clients, as follows:
Windows Server 2008: In Username and Group/Domain name, specify the operating system user account under which you want the backup session to run (for example, the user name
Administrator , domain DP ).
dba .
Ensure that this user has been added to the Data Protector admin or operator user group, has the
SAP R/3 backup rights, and has been set up for the Data Protector Inet service user impersonation. This user becomes the backup owner.
For details on setting accounts for the Inet service user impersonation, see the HPE Data
Protector Help index: “Inet user impersonation”.
Data Protector (10.00) Page 140 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Specifying an SAP R/3 system and Oracle instance
Click Next.
6. In the Configure SAP dialog box, specify the pathname of the Oracle Server home directory and
SAP R/3 data home directory. If you leave the fields empty, the default
ORACLE_HOME
directory is used.
Under Oracle login information to target database, specify the following: l
For the database authentication mode, specify Username, Password, and Service.
l
For the local operating system authentication mode, leave Username, Password, and Service empty.
l
For the remote operating system authentication mode, specify only Service (leave Username and Password empty).
The following are the option descriptions: l
Username and Password: Specify the user name and password of the Oracle database user
account described in Configuring user accounts, on page 133 .
l
Service : Specify the Oracle service name.
In Backup and restore executables directory, specify the pathname of the directory in which the SAP backup utilities reside. By default, the utilities reside in:
Windows systems: \\SAP_system\sapmnt\ORACLE_SID\sys\exe\run
UNIX systems: /usr/sap/ORACLE_SID/SYS/exe/run
Data Protector (10.00) Page 141 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Configuring an SAP R/3 database on a UNIX system (operating system authentication mode)
Configuring an SAP R/3 database on a Windows system (database authentication mode)
Data Protector (10.00) Page 142 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Click OK.
7. The SAP R/3 database is configured. Exit the GUI or proceed with creating the backup
Using the Data Protector CLI
1. Log in to the SAP R/3 system using the Oracle operating system user account.
2. At the command prompt, change current directory to the following directory:
Windows systems: Data_Protector_home\bin
HP-UX, Solaris systems:
/opt/omni/lbin
3. Execute: util_sap.exe -CONFIG ORACLE_SIDORACLE_HOMEtargetdb_connection_stringSAPTOOLS_
DIR [SAPDATA_HOME][SQL_PATH]
Parameter description
ORACLE_SID
Oracle instance name.
ORACLE_HOME
Pathname of the Oracle Server home directory.
Data Protector (10.00) Page 143 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
targetdb_ connection_ string
This argument value determines the authentication mode used for accessing the Oracle database: l
To select the database authentication mode, specify the login information to the target database in the format user_name/password@Oracle_
service
.
l
To select the local operating system authentication mode, specify only the character
/
.
l
To select the remote operating system authentication mode, specify the login information to the target database in the format
/@Oracle_service
.
SAPTOOLS_DIR
Pathname of the directory that stores the SAP backup utilities.
SAPDATA_HOME
Pathname of the directory where the SAP R/3 data files are installed. By default, this parameter is set to
ORACLE_HOME
.
The message *RETVAL*0 indicates successful configuration.
Handling errors
If you receive the message
*RETVAL*error_number
where error_number is different than zero, an error occurred.
To get the error description, execute:
Windows systems:
Data_Protector_home\bin\omnigetmsg 12 error_number
HP-UX systems:
/opt/omni/lbin/omnigetmsg 12 error_number
TIP:
To get a list of Oracle instances that are used by the SAP R/3 application, execute: util_sap.exe -APP
To get a list of tablespaces of an Oracle instance, execute: util_sap.exe -OBJS0ORACLE_SID
To get a list of database files of a tablespace, execute: util_sap.exe -OBJS1ORACLE_SID TABLESPACE
Checking the configuration
You can check the configuration of an SAP R/3 database after you have created at least one backup specification for this database. Use the Data Protector GUI or CLI.
Data Protector (10.00) Page 144 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Using the Data Protector GUI
1. In the Context List, select Backup.
2. In the Scoping Pane, expand Backup Specifications and then SAP R/3. Click the backup specification to display the Oracle instance to be checked.
3. Right-click the Oracle instance and click Check configuration.
Checking the SAP R/3 configuration
Using the Data Protector CLI
Log in to the SAP R/3 system as the Oracle OS user and execute: util_sap.exe -CHKCONF ORACLE_SID where
ORACLE_SID
is the name of the Oracle instance.
A successful configuration check displays the message
*RETVAL*0
.
If you receive the message *RETVAL*error_number where error_number is different than zero, an error
occurred. On how to get the error description, see Handling errors, on the previous page
.
To check if the SAP R/3 configuration is suitable for instant recovery, execute: util_sap.exe -CHKCONF_IR ORACLE_SID [-verbose]
Data Protector (10.00) Page 145 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
The
-verbose
option creates a file with a list of control files and redo log files that are on the same source volumes as the database files. If this list in not empty, a warning is displayed, stating that instant recovery is impossible.
Configuring the SAP R/3 parameter file
To configure the integration, you need to set some parameters in the SAP R/3 parameter file on both the application system and backup system. The file template is located in:
UNIX systems: ORACLE_HOME/dbs/initORACLE_SID.sap
Windows systems: ORACLE_HOME\database\initORACLE_SID.sap
SAP parameter file settings
Parameter
split_cmd primary_db
Value/Description
On the application system:
UNIX systems:
"/opt/omni/lbin/ob2smbsplit $"
Windows systems:
"Data_Protector_home\bin\ob2smbsplit $"
On the backup system, you do not need to set the parameter.
BRBACKUP uses this parameter to trigger the replica creation. At run time, the optional sign “ $ ” is replaced with the name of the text file containing the names of files to be backed up.
Windows systems: If the pathname contains spaces, use Windows short names instead.
On the application system:
LOCAL
On the backup system: name of the service used for connecting to the
Oracle database.
This parameter defines the service name of the Oracle database to link the backup system to the application system.
Backup
The integration provides online and offline database backups of the following types: l
ZDB to disk l
ZDB to tape l
ZDB to disk+tape
To configure a backup, create a ZDB backup specification.
Data Protector (10.00) Page 146 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Archived logs can only be backed up together with the database in a ZDB to disk+tape, ZDB to tape, or non-ZDB (standard backup) session. If you try to back up archived logs in a ZDB to disk session, or archived logs only in a ZDB session, the session fails.
What is backed up
Selected items Backed up files
ARCHIVELOGS
l offline (archived) redo logs l control files
DATABASE or individual tablespaces l data files l control files l
SAP R/3 logs and parameter files l online redo logs (only during offline backups)
You can specify SAP R/3 backup options in two different ways: l
Using the BRBACKUP options.
l
Using the SAP parameter file.
NOTE:
The BRBACKUP options override the settings in the SAP parameter file.
You can specify BRBACKUP options when you create a backup specification. If no options are specified, the SAP R/3 application refers to the current settings in the SAP parameter file. In such a case, before running a backup, ensure that the SAP parameter file is correctly configured.
TIP:
When you create a backup specification, select a backup template that already contains the desired BRBACKUP options.
Considerations
l
Before you start a backup, ensure that the SAP R/3 database is in the open or shutdown mode.
l
Before you start a backup, ensure the primary_db parameter is set to LOCAL in the SAP R/3
parameter file. For more information on setting the SAP R/3 parameter file, see Configuring the SAP
R/3 parameter file , on the previous page
.
l
ZDB, restore, and instant recovery sessions that use the same source volume on the application system cannot run simultaneously.
l
You cannot start a ZDB to disk session if another session is backing up the archived logs, even if the Oracle data files and the archived logs reside on different source volumes.
l
P9000 XP Array: If the LVM Mirroring configuration is used, Data Protector displays a warning during a backup because the volume group source volumes on the application system do not have their BC pairs assigned. The message should be ignored.
Data Protector (10.00) Page 147 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration l
ZDB to disk: Archived logs cannot be backed up. To back up archived logs, create a non-ZDB
(standard) SAP R/3 backup specification. For information, see the HPE Data Protector Integration
Guide.
l
Configurable backup modes are supported only by using templates.
l
By default, Data Protector supports all BRTOOL options, except -a and -b . To enable support for -a and -b , set the OB2BRTNOSECUomnirc option to 1 . On how to set the omnirc options, see the HPE
Data Protector Help index: “omnirc options”.
Creating backup specifications
Create a backup specification using the Data Protector Manager.
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click SAP R/3, and click Add
Backup.
3. In the Create New Backup dialog box, select a template ( Selecting a template, below ) and click
OK.
Selecting a template
Backup templates available for zero downtime backup
Data Protector (10.00) Page 148 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Blank SAP Backup
Brbackup_offline_mirror
No predefined options.
Used for an offline ZDB (split mirror or snapshot backup) using splitint . The database is stopped during the creation of a replica. The database is offline for a shorter period of time than if Brbackup_SMB_
Offline was used, but splitint
must be supported by
BRTOOLS.
Brbackup_online_mirror
Used for an online ZDB (split mirror or snapshot backup) using splitint . The database is active during the creation of a replica. The database is in the backup mode for a shorter period of time than if Brbackup_
SMB_Online was used, but splitint
must be supported by BRTOOLS.
Brbackup_SMB_Offline
Brbackup_SMB_Online
Used for an offline ZDB (split mirror or snapshot backup). The database is stopped during the creation of a replica.
Used for an online ZDB (split mirror or snapshot backup). The database is active during the creation of a replica.
From the Backup type drop-down list, select Snapshot or split mirror backup.
From the Sub type drop-down list, select the appropriate disk array agent. The agent must be installed on the application system and the backup system.
4. In Application system, select the SAP R/3 client to be backed up. In cluster environments, select the virtual server.
In Backup system, select the backup system.
Select other disk array-specific backup options (see EMC backup options, on the next page for
EMC, P9000 XP Array backup options, on the next page
for P9000 XP Array, P6000 EVA Array backup options, on page 151
for P6000 EVA Array, NetApp Storage backup options, on page 152
for NetApp Storage). For detailed information on the backup options, press F1.
Data Protector (10.00) Page 149 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
EMC backup options
P9000 XP Array specifics
To enable instant recovery, leave the Track the replica for instant recovery option selected. It is not possible to run instant recovery with Data Protector if this option is cleared.
P9000 XP Array backup options
Data Protector (10.00) Page 150 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
P6000 EVA Array specifics
To enable instant recovery, select the Track the replica for instant recovery option.
P6000 EVA Array backup options
P10000 3PAR Array specifics
P10000 3PAR Array backup options
Data Protector (10.00) Page 151 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
NetApp Storage backup options
Click Next.
5. In Application database, select the Oracle instance ( ORACLE_SID ) to be backed up.
Specify the User and group/domain options, which are available on UNIX and Windows Server
2008 clients, as follows:
Windows Server 2008: In Username and Group/Domain name, specify the operating system user account under which you want the backup session to run (for example, the user name
Administrator
, domain
DP
).
dba
.
Ensure that this user has been added to the Data Protector admin or operator user group, has the
SAP R/3 backup rights, and has been set up for the Data Protector Inet service user impersonation. This user becomes the backup owner.
For details on setting accounts for the Inet service user impersonation, see the HPE Data
Protector Help index: “Inet user impersonation”.
Click Next.
6. If the SAP R/3 database is not configured yet for use with Data Protector, the Configure SAP
dialog box is displayed. Configure it as described in Configuring SAP R/3 databases, on page 139
.
7. Select SAP R/3 objects to be backed up. You can select individual tablespaces, data files, or archived logs.
NOTE:
If you plan to do instant recovery, select the whole DATABASE item.
Data Protector (10.00) Page 152 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Selecting backup objects
Click Next.
8. Select devices to use for the backup.
To specify device options, right-click the device and click Properties. Specify the number of parallel backup streams in the Concurrency tab and the media pool.
NOTE:
Parallelism (the number of streams your SAP R/3 database is backed up with) is set automatically. If load balancing is used, the number of streams equals the sum of concurrencies of the selected devices.
Click Next.
Application-specific options
Data Protector (10.00) Page 153 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Click Next.
10. Click Save As to save the backup specification, specify a name and a backup specification group. Optionally, click Save and Schedule to save, and then schedule the backup specification.
For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE
Data Protector Administrator's Guide.
TIP:
SAP R/3 backup options
Option Description
Log file
BR Backup
If you want to create a backint log file during backup, specify a pathname for the file. By default, this file is not created because Data
Protector stores all relevant information about backup sessions in the database.
Specifies BRBACKUP options.
For example, for online backup using the splitint
interface, type
-t online_mirror . If splitint is not supported by BRTOOLS, type -t online_split .
To run BRBACKUP under a different Oracle database user than the
Data Protector (10.00) Page 154 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Option
Backup Objects
BR Archive
Balancing: By Load
Balancing: By Time
Balancing: Manual
Balancing: None
Pre-exec , Post-exec
Description
one specified during the configuration, type -u user_name .
Lists BRBACKUP options passed by omnisap.exe
. The list is displayed after you save the backup specification.
Specifies BRARCHIVE options.
Not applicable for ZDB to disk.
Groups files into subsets of approximately equal sizes. The subsets are then backed up concurrently by Data Protector sapback programs.
If your backup devices use hardware compression, the sizes of the original and backed up files differ. To inform Data Protector of this, specify the original sizes of the backed up files in the compression
section of the Data Protector SAP R/3 configuration file. See Data
Protector SAP R/3 configuration file , on page 127 .
Groups files into subsets that are backed up in approximately equal periods of time. The duration depends on the file types, speed of the backup devices, and external influences (such as mount prompts).
This option is best for environments with large libraries of the same quality. The subsets are backed up concurrently by Data Protector sapback programs. Data Protector automatically stores backup speed information in the speed
section of the Data Protector SAP R/3 configuration file. It uses this information to optimize the backup time.
This type of balancing may lead to non-optimal grouping of files in the case of an online backup or if the speed of backup devices varies significantly.
Groups files into subsets as specified in the manual balancing section of the Data Protector SAP R/3 configuration file. For more information,
see Manual balancing, on page 160
.
Not applicable for ZDB to disk.
No balancing is used. The files are backed up in the same order as they are listed in the internal Oracle database structure. To check the order, use the Oracle Server Manager SQL command: select * from
dba_data_files
The command specified here is started by omnisap.exe
on the SAP
R/3 system before the backup ( pre-exec ) or after it ( post-exec ). Do not use double quotes. Provide only the name. The command must reside in the directory:
Windows systems: Data_Protector_home\bin
Data Protector (10.00) Page 155 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Option
Backup mode
Use default RMAN channels
Objects outside database
Description
HP-UX, Solaris: /opt/omni/bin
Not applicable for ZDB.
Not applicable for ZDB.
Specifies non-database files of the Oracle SAP R/3 environment to be saved.
Save these files in a separate backup session.
NOTE:
The total number of sapback processes started in one session using Data Protector is limited to
256.
Modifying backup specifications
To modify your backup specification, click its name in the Scoping Pane of the Backup context, then click the appropriate tab, and apply the changes.
Scheduling backup sessions
You can run unattended backups at specific times or periodically. For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE Data Protector Administrator's
Guide.
Previewing backup sessions
Preview the backup session to test it. You can use the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications and then SAP R/3. Right-click the backup specification you want to preview and click Preview Backup.
3. Specify Backup type and Network load. Click OK.
The message
Session completed successfully
is displayed at the end of a successful preview.
Using the Data Protector CLI
Execute: omnib -sap_list backup_specification_name -test_bar
Data Protector (10.00) Page 156 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
What happens during the preview?
The omnisap.exe
command is started, which starts the Data Protector testbar command to test the following: l
The syntax of the backup specification l
If devices are correctly specified l
If the necessary media are in the devices
Starting backup sessions
Interactive backups are run on demand. They are useful for urgent backups or restarting failed backups.
Backup methods
Start a backup of SAP R/3 objects in any of the following ways: l
Using the Data Protector GUI.
l
Using the Data Protector CLI.
l
Using the SAP BR*Tools.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, and then SAP R/3. Right-click the backup specification you want to use and click Start Backup.
3. Specify Click OK.
NOTE:
Only the Full backup type is supported.
ZDB to disk, ZDB to disk+tape: Specify the Split mirror/snapshot backup option.
The message Session completed successfully is displayed at the end of a successful backup session.
Using the Data Protector CLI
Execute:
ZDB to tape, ZDB to disk+tape:
omnib -sap_list backup_specification_name
ZDB to disk:
omnib -sap_list backup_specification_name -disk_only
For details, see the omnib man page or the HPE Data Protector Command Line Interface Reference.
Data Protector (10.00) Page 157 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Using the SAP BRTOOLS
1. Log in to the SAP R/3 backup system or SAP R/3 application system as the Oracle OS user.
2. Export/set the following environment variables:
ORACLE_SID=SAP_
instance_name
ORACLE_HOME=Oracle_
software_home_ directory
[SAPBACKUP_
TYPE=OFFLINE]
SAPDATA_
HOME=database_
files_directory
SAPBACKUP=BRTOOLS_
logs_and_control_ file_copy_directory
Default is ONLINE .
SAPREORG=BRSPACE_
logs_directory
OB2BARLIST=backup_
specification_name
The backup specification is needed only to specify which Data
Protector devices should be used for backup. Other information from the backup specification, like SAP R/3 objects to be backed up or the
BRBACKUP options, is ignored and has to be specified manually at run time.
[OB2_3RD_PARTY_
BACKINT=1]
Specifies usage of a third-party backint tool to perform ZDB-todisk+tape backup sessions. After setting the variable, copy the backint you want to use to the SAP BRTOOLS directory.
[OB2BARHOSTNAME=
application_system_
name]
Required if you are logged in to the backup system. Optional if you want to specify a virtual server name in cluster environments.
[OB2BACKUPHOSTNAME=
backup_system_name]
Required if you are logged in to the application system.
OB2SMB=1 Specifies a ZDB session.
Specifies tracking replica for instant recovery.
[OB2SMBIR=1]
[ZDB_ORA_INCLUDE_
CF_OLF=1]
[OB2DISKONLY=1]
Required if you are logged in to the backup system.
Specifies a ZDB-to-disk session.
Data Protector (10.00) Page 158 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
If you are logged on to the backup system, ensure that the NLS_LANG environment variable is set to the same value as the
NLS_LANG
environment variable on the application system.
Alternatively, these variables can be specified in the backint parameter file. If this is required, the location of the file must be specified in the SAP configuration file using the
util_par_file
parameter: util_par_file = path\filename
If you do not supply the path, the system searches for the parameter file in the directory:
Windows systems: SAPDATA_HOME\database
UNIX systems:
ORACLE_HOME/dbs
3. Execute the BRBACKUP command. The command syntax depends on whether you are logged in to the application system or backup system:
Application system: brbackup -t {online_split | offline_split | online_mirror | \ offline_mirror}
[–q split] -d \ util_file -m all -c -u user/password
Backup system: brbackup -t {online_mirror | offline_mirror} [–q split] -d util_file -m all -c
-u user/password
The -q split option is required if OB2DISKONLY is set to 1 .
Configuring SAP compliant ZDB sessions
SAP R/3 standards recommend that, in ZDB sessions that use the splitint backup interface,
BRBACKUP is started on the backup system and not on the application system. You can configure
Data Protector to comply with these standards by setting the Data Protector
OB2_MIRROR_COMP environment variable to 1 . The variable is saved in the Data Protector SAP R/3 instance configuration file. Consequently, in all splitint ZDB sessions for this SAP R/3 instance, BRBACKUP will be started on the backup system. By default, BRBACKUP is started on the application system.
Set the OB2_MIRROR_COMP environment variable using the Data Protector GUI or CLI.
NOTE:
If no backup specification for the related SAP R/3 instance exists, you cannot use the Data
Protector CLI to set the OB2_MIRROR_COMP variable.
Using the Data Protector GUI
You can set the OB2_MIRROR_COMP variable when you create a backup specification or modify an existing one:
1. Proceed to the Source page of the backup specification.
NOTE:
In environments in which the control file and datafiles reside on the same source disk, Data
Protector does not let you proceed to the Source page if the Track the replica for instant
recovery option is selected. Specifically, the Data Protector instant recovery check fails.
Data Protector (10.00) Page 159 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
In such a case, clear the option first. You can select the option later if needed, when the
OB2_MIRROR_COMP variable is already set and, consequently, the instant recovery check is no longer performed.
Right click the SAP R/3 instance at the top and click Set Environment Variables.
2. In the Advanced dialog box, set OB2_MIRROR_COMP to 1
. See Setting environment variables, below
.
Click OK.
Setting environment variables
Using the Data Protector CLI
Execute the following command: util_cmd –putopt SAP instance_name OB2_MIRROR_COMP 1 –sublist Environment
Manual balancing
Manual balancing means that you manually group files into subsets, which are then backed up in parallel. To group files into subsets, add the manual_balance section to the Data Protector SAP R/3 configuration file as described in the following example.
Data Protector (10.00) Page 160 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Example
Suppose that we have a backup specification named
SAP-R3
with the following files to be backed up: fileA
, fileB
, fileC
, fileD
. To group the files into three subsets (
0={fileA, fileC}
,
1={fileB}
,
2=
{fileD} ), add the following lines to the Data Protector SAP R/3 configuration file: manual_balance={
SAP-R3={ fileA=0; fileB=1; fileC=0;fileD=2;}}
When you group files into subsets, consider the following: l
Use only one file from the same hard disk at a time.
l
The number of files in a subset must be equal to or smaller than the sum of the concurrencies of all devices specified for backup.
l
If the backup specification contains files that are not allocated to any subset, Data Protector automatically adds these files to the list of files to be backed up using the load balancing principle.
Before the backup, this list is logged in:
Windows systems:
SAPDATA_HOME\sapbackup\*.lst
UNIX systems: ORACLE_HOME/sapbackup/.*.lst
Restore
You can restore SAP R/3 objects using any of the following methods: l
Standard restore: Data is restored from backup media created in ZDB to tape, ZDB to disk+tape,
and non-ZDB (standard backup) sessions. See Standard restore, on page 163 .
l
Instant recovery: Data is restored from a replica created in online ZDB–to–disk or ZDB–to–
disk+tape sessions. See Instant recovery , on page 166
.
After the restore, you can recover the database to a specific point in time using the SAP BRTOOLS interface. Instant recovery method enables you to restore and recover the database within the same session. However, you can only restore (and recover) the whole database. To restore only a part of the database or the archived logs, use the standard restore method.
shows which restore methods are available, depending on the backup session you restore from.
SAP recovery methods
Data Protector (10.00) Page 161 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Disk array Backup types Recovery of the whole database
Until now
Restore
Recovery of a part of the database
To a point in time, logseq/thread or SCN number
Restore Restore
P9000 XP,
P6000 EVA,
EMC, NetApp
Storage
P9000 XP,
P6000 EVA
ZDB to tape - online
ZDB to tape - offline
ZDB to disk - online
Restore Restore Restore
ZDB to disk - offline
ZDB to disk+tape - online
l
Instant recovery + database recovery or l
Restore
Restore
ZDB to disk+tape - offline
Instant recovery
+ database recovery
Instant recovery
+ database recovery
N/A
N/A N/A N/A l
Instant recovery + database recovery l or
Restore
Restore
Restore
Restore
Legend
Restore
Instant recovery + database recovery
N/A
You can do a standard restore from the Data Protector media using the
Data Protector GUI or the SAP BRTOOLS. After the restore, you can recover the database using the SAP BRTOOLS.
You can do an instant recovery. You can include database recovery in the instant recovery session or do it afterwards, using the SAP BRTOOLS.
Not available.
Data Protector (10.00) Page 162 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Considerations
l
SAP R/3 tablespaces located on raw partitions cannot be restored using the Data Protector GUI.
Workaround: Use SAP restore commands (for example, brrestore ).
l
If your Oracle database is localized, you may need to set the appropriate Data Protector encoding
before you start a restore. For details, see Localized SAP R/3 objects, on page 171
.
l
Restore preview is not supported.
Standard restore
Restore SAP R/3 objects using the Data Protector Manager.
1. In the Context List, click Restore.
2. In the Scoping Pane, expand SAP R/3, expand the client (backup system) from which the data was backed up, and then click the Oracle instance you want to restore.
3. In the Source page, select SAP R/3 files to be restored.
To restore a file under a different name or to a different directory, right-click the file and click
Restore As/Into.
To restore a file from a specific backup session, right-click the file and click Restore Version.
Selecting objects for restore
4. In the Destination tab, select the client to restore to (Target client). By default, this is the
application system. See Selecting the target client, below .
For details on options, press F1.
Selecting the target client
Data Protector (10.00) Page 163 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
5. In the Options page, set the restore options. For information, press F1.
6. In the Devices page, select the devices to be used for the restore.
For more information on how to select devices for a restore, see the HPE Data Protector Help index: “restore, selecting devices for”.
7. Click Restore.
8. In the Start Restore Session dialog box, click Next.
9. Specify Report level and Network load.
NOTE:
Select Display statistical information to view the restore profile messages in the session output.
10. EMC and P9000 XP Array: This step is relevant only if you have both the EMC Symmetrix Agent and HPEP9000 XP Agent components installed on the application system.
EMC: Select EMC Symmetrix restore .
P9000 XP Array: Select
P9000 XP restore
. See Selecting the P9000 XP restore, on the next page
.
Data Protector (10.00) Page 164 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Selecting the P9000 XP restore
Click Next.
11. EMC and P9000 XP Array: From the EMC Symmetrix mode or Mirror mode drop-down list, select Disabled. This sets the restore from backup media to the application system directly. See
EMC – Selecting restore to the application system directly, below
.
EMC – Selecting restore to the application system directly
Data Protector (10.00) Page 165 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
P9000 XP Array – Selecting restore to the application system directly
12. Click Finish to start the restore.
The statistics of the restore session, along with the message Session completed successfully is displayed at the end of the session output.
Instant recovery
For general information on instant recovery, see the HPE Data Protector Concepts Guide and HPE
Data Protector Zero Downtime Backup Administrator's Guide. For information on instant recovery in cluster environment (Cluster File System (CFS), HPE Serviceguard, and Microsoft Cluster Server), see the HPE Data Protector Zero Downtime Backup Administrator's Guide.
You can start an instant recovery using the Data Protector GUI or CLI.
Considerations
l
The database recovery part is performed after the instant recovery procedure. During database recovery, archive log backups performed after the ZDB are restored from tape by the SAP BR*Tools utilities. If selected, the logs are reset and the database is opened.
l
Instant recovery using the Data Protector GUI
To perform an instant recovery:
1. Shut down the Oracle database using sqlplus :
For example:
Data Protector (10.00) Page 166 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration sqlplus sqlplus> shutdown immediate sqlplus> exit
2. In the Data Protector Manager Context List, select Instant Recovery.
3. Expand SAP R/3 and select the ZDB-to-disk or ZDB-to-disk+tape session from which you want to perform the restore.
4. In the Source tab, select the objects to recover. Only whole databases can be selected.
For HPE P9000 XP Disk Array Family, it is recommended to leave the Keep the replica after the
restore option selected to enable a restart of an instant recovery session. The option is selected by default, except for an offline backup where the database was in NOARCHIVELOG mode during the backup. With HPE P6000 EVA Disk Array Family, replica is kept on the disk array only if the Copy replica data to the source location is selected.
Set other HPE P6000 EVA Disk Array Family or HPE P9000 XP Disk Array Family options. For details, press F1.
SAP R/3 source options
5. At this point, you can decide whether to perform a database recovery immediately after an instant recovery or not: l
To perform only an instant recovery, click Restore. l
To automatically perform a database recovery after an instant recovery, select the recovery
Data Protector (10.00) Page 167 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
options. For details, see Database recovery options , below
.
Click Restore.
Data Protector recovers the database after performing instant recovery by switching the database to mount state, restoring the necessary archive redo logs from tape, and applying the redo logs.
Database recovery options
User name
(UNIX systems only)
Specifies the user name under which the instant recovery is performed. The user needs to be a member of the DBA group.
User group
(UNIX systems only)
Specifies the user group the user in the User name field belongs to.
NOTE:
The User name and the User group must be the same as defined in the backup ownership.
Recovery
Enables database recovery after instant recovery.
Recover until
The options in this drop-down list enables you to specify to which point in time you would like the recovery to be performed.
The following options are available:
Now : All existing archive logs are applied.
Selected time : Only archive logs until the specified time are applied.
Selected logseq/ thread number : Specifies an incomplete recovery. Only archive logs with a lower or equal number than the specified log sequence or thread number are applied.
Selected SCN number : Only archive logs until the specified SCN number are applied.
Open database after recovery
Opens the database after the recovery was performed.
Reset logs
Resets the archive logs after the database is opened. This option is by default not selected if the Recover until option is set to Now.
The following are Oracle recommendations on when to reset the logs:
Always reset the logs: l
After an incomplete recovery, that is, if not all archive redo logs are applied.
l
If a backup of the control file is used for recovery.
Do not reset the logs: l
After a complete recovery, when the control file is not used. l
If the archive logs are used for a standby database. However, if you must reset the archive logs, recreate the standby database.
Data Protector (10.00) Page 168 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
SAP R/3 recovery options
Instant recovery using the Data Protector CLI
Execute: omnir
–host ClientName
-session SessionID
-instant_restore
[P9000_DISK_ARRAY_XP_OPTIONS | P6000_ENTERPRISE_VIRTUAL_ARRAY_OPTIONS]
-sap
-user UserName -group GroupName
-recover {now | time MM/DD/YY hh:mm:ss | logseq LogSeqNumber thread ThreadNumber |
SCN Number} [-open [-resetlogs]]
-appname ApplicationDatabaseName
The order of options is important. On Windows clients, the user name and group name options are not required. For a detailed description of the options, see the HPE Data Protector Command Line
Interface Reference.
Data Protector (10.00) Page 169 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Instant recovery from replicas containing the control file
During an instant recovery from a replica that contains the control file, the current control file and, possibly, online redo logs get overwritten. Therefore, before you start the session, copy the current control file and online redo logs to a safe location to be able to do a database recovery afterwards.
A replica contains the control file if it is created in any of the following sessions: l
Online ZDB session with the omnirc option ZDB_ORA_INCLUDE_CF_OLF set to 1 l
Online SAP compliant ZDB session l
Offline ZDB session (any configuration)
NOTE:
An offline ZDB session also contains online redo logs. You can use such a session to restore the SAP R/3 database to a point in time at which the backup was performed. In this case, you do not need to follow the steps below.
To restore and recover the database:
1. Copy the current control files and online redo logs to a safe location.
2. Perform instant recovery (without database recovery). Use the Data Protector GUI or CLI.
3. Copy the current control files and online redo logs back to their original location.
4. Mount the target database.
5. Restore missing archived redo logs required for database recovery.
Example:
# sqlplus
user/password@net_service_name
SQL> select SEQUENCE#, NAME from V$ARCHIVED_LOG where
(NEXT_TIME>to_date('2010/10/03','YYY/MM/DD') and (FIRST_CHANGE#<='1000');
# brrestore -a log_no,... -d util_file -c force -u user/password
6. Recover the target database.
Example:
# rman target user/password@net_service_name
RMAN> run{
2> allocate channel dbrec type disk;
3> recover database until scn 1000;
4> release channel dbrec;
5> }
Restoring using another device
You can perform a restore using a device other than that used for the backup.
Data Protector (10.00) Page 170 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Using the Data Protector GUI
For information on how to select another device for a restore using the Data Protector GUI, see the
HPE Data Protector Help index: “restore, selecting devices for”.
Using the Data Protector CLI or SAP commands
If you are restoring using the Data Protector CLI or SAP R/3 commands, specify the new device in the file:
Windows systems: Data_Protector_program_data\Config\Server\cell\restoredev
UNIX systems: /etc/opt/omni/server/cell/restoredev
Use the format:
"
DEV 1
" "
DEV 2
" where DEV 1 is the original device and DEV 2 the new device.
IMPORTANT:
Delete this file after use.
On Windows systems, use the Unicode format for the file.
Localized SAP R/3 objects
Oracle Server uses its own encoding, which may differ from the encoding used by the filesystem. In the
Backup context, Data Protector displays the logical structure of the Oracle database (with Oracle names) and in the Restore context, the filesystem structure of the Oracle database. Therefore, to display non-ASCII characters correctly, ensure that the Data Protector encoding matches with the
Oracle Server encoding during backup and with the filesystem encoding during restore. However, the incorrect display does not impact the restore.
Windows systems: If the current values of DBCS and the default Windows character set for non-
Unicode programs do not match, problems arise. See Restore problems, on page 192 .
UNIX systems: To be able to switch between the Data Protector encodings, start the GUI in UTF-8 locale.
Monitoring sessions
You can monitor currently running sessions in the Data Protector GUI. When you run an interactive backup or a restore session, a monitor window shows you the progress of the session. Closing the GUI does not affect the session.
You can also monitor sessions from any Data Protector client with the User Interface component installed, using the Monitor context.
On how to monitor a session, see the HPE Data Protector Help index: “viewing currently running sessions”.
Data Protector (10.00) Page 171 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
System messages generated during backups are sent to both the SAP R/3 and the Data Protector monitor. However, mount requests are sent only to the Data Protector monitor.
Troubleshooting
This section lists general checks and verifications plus problems you might encounter when using the
Data Protector SAP R/3 integration.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
For general ZDB, restore, and instant recovery related troubleshooting, see the troubleshooting sections in the HPE Data Protector Zero Downtime Backup Administrator's Guide.
See also the troubleshooting section in the SAP R/3 chapter of the HPE Data Protector Integration
Guide.
Before you begin
l
Ensure that the latest official Data Protector patches are installed. See the HPE Data Protector Help index: “patches” on how to verify this.
l
See the HPE Data Protector Product Announcements, Software Notes, and References for general
Data Protector limitations, as well as recognized issues and workarounds.
l
For an up-to-date list of supported versions, platforms, and other information, see the latest support matrices at https://softwaresupport.hpe.com/manuals .
General troubleshooting
Problem
Configuration fails due to a database operation failure
During configuration of an SAP R/3 database, Data Protector reports the following error:
Integration cannot be configured.
The database reported error while performing requested operation.
Action
Review user group membership for the user account which is used in Oracle database access
authentication. For details, see Configuring user accounts, on page 133
.
Prerequisites on the SAP side of the integration
The following verification steps must be performed in order to verify that SAP is installed as required for the integration to work. These steps do not include Data Protector components.
1. Verify backup directly to disk as follows: brbackup -d disk -u user/password
Data Protector (10.00) Page 172 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
If this fails, check the error messages and resolve possible problems before you continue.
2. Verify restore directly to disk as follows: brrestore -d disk -u user/password
If this fails, check the error messages and resolve possible problems before you continue.
3. If you are running backups in RMAN mode, verify backup and restore directly to disk
using Recovery Manager channel type disk as follows:
a. You must define the parameter init in the initialization file init ORACLE_SID.
ora . Execute the following commands: brrestore -d pipe -u user/password -t online -m all brrestore -d disk -u user/password
b. If this fails, see the SAP Online Help to learn how to execute backup and restore directly to disk using the SAP backup utility.
Check the error message and resolve these problems before you continue.
4. Verify that the SAP backup tools correctly start backint (which is provided by Data
Protector):
Move the original backint and create a test script namedbackint.bat in the directory where the SAP backup utility resides, with the following entries: echo "Test backint called as follows:" echo "%0%1%2%3%4%5%6%7%8%9" exit
Then start the following commands: brbackup -t offline -d util_file -u user/password -c
If you receive backint arguments, this means that SAP is properly configured for backup using backint; otherwise you have to reconfigure SAP.
See Configuring SAP R/3 databases, on page 139
.
Configuration problems
IMPORTANT:
The procedure described in the previous sections must be performed before you start checking the Data Protector configuration.
1. Verify that the Data Protector software has been installed properly.
For details, see the HPE Data Protector Installation Guide.
2. Perform a filesystem backup of the SAP Database Server.
Perform a filesystem backup of the SAP Database Server system so that you can eliminate any potential communication problems between the SAP Database Server and the Data Protector Cell
Manager system.
Do not start troubleshooting an online database backup unless you have successfully completed a filesystem backup of the SAP Database Server system.
See the HPE Data Protector Help index “standard backup procedure” for details about how to do a filesystem backup.
Data Protector (10.00) Page 173 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
3. If the SAP backup utilities are installed in a shared directory, then the inet startup
parameter must be specified as described in
If you use the command line to start the Data
Protector commands, verify the inet startup parameters., below , or the Windows
permissions must be set correctly.
Execute the following command (if you use the default directory): dir \\client_name\sapmnt\ORACLE_SID\SYS\exe\run\brbackup or dir \\client_name\SAPEXE\brbackup
If this fails, set the inet
startup parameters, or set the correct permissions to access a Windows network directory.
4. If you use the command line to start the Data Protector commands, verify the inet startup
parameters.
Check the Data Protector Inet service startup parameters on the SAP Database Server system. Proceed as follows:
a. In the Control Panel, go to Administrative Tools, Services.
b. Select Data Protector Inet.
In the Services window, select Data Protector Inet, Startup.
The service must run under a specified user account. Make sure that the same user is also added to the Data Protector admin
user group.
Checking the Inet start-up parameters
Data Protector (10.00) Page 174 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
5. Examine the environment variables.
If you need to export some variables before starting the Oracle Server Manager, TNS listener, or other Oracle utility, these variables must be defined in the Environment section of the Data
6. Examine system errors.
System errors are reported in the debug.log
file on the SAP Server.
Problem
Configuration fails due to a script failure
During configuration of an SAP R/3 database, Data Protector reports the following error:
Integration cannot be configured.
Script failed. Cannot get information from remote host.
Action
Check the environment settings and ensure Data Protector Inet is running under a user account which
has the required privileges. For details, see Before you begin, on page 133 .
Backup problems
At this stage, you should have performed all the verification steps described in the previous sections. If backup still fails, proceed as follows:
1. Check your SAP Server configuration:
To check the configuration, start the following command on the SAP Server system:
Data_Protector_home\bin\util_sap.exe -CHKCONF ORACLE_SID
The message *RETVAL*0 indicates successful configuration.
2. Verify Data Protector internal data transfer using the testbar2 utility.
Before you run the testbar2 utility, verify that the Cell Manager name is correctly defined on the
SAP Database Server. in In the default Data Protector client configuration directory, check the cell_server
file, which contains the name of the Cell Manager system. Then execute the following command:
Data_Protector_home\bin\testbar2 -type:SAP -appname:ORACLE_SID -bar:backup_
specification_name -perform:backup
Examine the errors reported by the testbar2 utility by clicking the Details button in the Data
Protector Monitor context.
If the messages indicate problems concerning the Data Protector side of the integration, create an
SAP backup specification to back up to a null or file device. If the backup succeeds, the problem may be related to the backup devices. For instructions on troubleshooting devices, see the HPE
Data Protector Troubleshooting Guide. If the test fails again, call support.
3. Verify the backup using backint export OB2BARLIST=barlist_name export OB2APPNAME=ORACLE_SID
Data Protector (10.00) Page 175 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Data_Protector_home\bin\backint.exe -f backup -t file -u ORACLE_SID -i input_
file
where
input_file
is a file with a list of full pathnames for backup.
Backint anticipates a list of files in the following format:
pathName_1pathName_2pathName_3
Problem
Backup fails with “Connect to database instance failed”
If you start a backup while the database instance is in the unmount or mount mode, the session fails with a message similar to the following:
BR0301E SQL error -1033 at location BrDbConnect-2
ORA-01033: ORACLE initialization or shutdown in progress
BR0310E Connect to database instance HOOHOO failed
Action
Before you start a backup, ensure that the database instance is in the open or shutdown mode.
Restore problems
At this stage, you should have performed all the verification steps described in the previous sections.
After this, proceed as follows:
1. Verify that a backup object exists on the backup media and in the IDB:
This can be done by executing the command omnidb -sap "object_name" -session "Session_ID" -media on the SAP Database Server system.
The output of the command lists detailed information about the specified backup object, session
IDs of the backup sessions containing this object, and a list of the media used.
For detailed syntax of the omnidb
command, execute: omnidb -help
You can also do this using the SAP tools:
Use backint, so that SAP tools also use this command to query:
Data_Protector_home\bin\backint.exe -f inquiry -u ORACLE_SID -i input_file where the specified
input_file
is queried.
If this fails, check if the backup session was performed successfully and if the query was started under the appropriate user account.
Backint anticipates a list of files of the following format:
backup_ID_1 pathName_1
[targetDirectory_1]
To retrieve the
backup_ID
numbers, enter the following command: echo #NULL #NULL | backint –f inquiry –u ORACLE_SID
Data Protector (10.00) Page 176 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration or, alternatively, you can just specify #NULL as
backup_ID_1
in the
input_file
. In this case, the latest backup session for the file is used for the restore.
2. Verify the restore using the Data Protector User Interface
This test is possible if the objects have been backed up by backint.
If this fails, check if the backup session was performed successfully and if the query was started under the appropriate user account.
3. Simulate a Restore Session
Once you know the information about the object to be restored, you can simulate a restore using the Data Protector testbar2
utility.
Before you execute testbar2 , verify that the Cell Manager name is correctly defined on the SAP
Database Server.
In the default Data Protector client configuration directory, check the cell_server file, which contains the name of the Cell Manager system.
Then, test the Data Protector internal data transfer using the testbar2 utility:
Data_Protector_home\bin\testbar2 -type:SAP
-appname:ORACLE_SID
-perform:restore
-object:object_name
-version:object_version
-bar:backup_specification_name
You should see only NORMAL messages displayed on your screen, otherwise examine the errors reported by the testbar2 utility by clicking the Details button in the Data Protector Monitor context.
4. Verify the restore using backint
Execute the following command:
Data_Protector_home\bin\backint.exe -f restore -u ORACLE_SID -i input_file where the contents of the
input_file
will be restored.
If this fails, check if the session was performed successfully and if the restore was started under the appropriate user account.
Backint anticipates a list of files in the following format: backup_ID_1pathName_1
[targetDirectory_1]backup_ID_2pathName_2 [targetDirectory_2]backup_ID_
3pathName_3 [targetDirectory_3]
To retrieve the
backup_ID
numbers, enter the following command: echo "#NULL #NULL" | backint –f inquiry –u ORACLE_SID
Problem
Restore sessions fail due to invalid characters in filenames
On Windows systems, where the Oracle Database Character Set (DBCS) is not set to the same value as the default Windows character set for non-Unicode programs, and where SAP tools are used to create Oracle datafiles, restore fails if the datafiles contain non-ASCII or non-Latin 1 characters.
Data Protector (10.00) Page 177 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Actions
Use any of the following solutions: l
For new Oracle installations, set the DBCS to UTF-8.
l
If you do not use other non-Unicode programs, set the language for non-Unicode programs to the same value as DBCS.
l
Do not use non-ASCII or non-Latin 1 characters for filenames.
Prerequisites on the SAP side of the integration
The following verification steps must be performed in order to verify that SAP is installed as required for the integration to work. These steps do not include Data Protector components.
1. Verify backup directly to disk as follows: brbackup -d disk -u user/password
If this fails, check the error messages and resolve possible problems before you continue.
2. Verify restore directly to disk as follows: brrestore -d disk -u user/password
If this fails, check the error messages and resolve possible problems before you continue.
3. If you are running backups in RMAN mode, verify backup and restore directly to disk
using Recovery Manager channel type disk as follows:
a. Re-link the Oracle Server with the Database Library provided by SAP (libobk.sl) .
For each RMAN channel, set the
SBT_LIBRARY
parameter to point to the libobk.sl
file.
IMPORTANT:
Before you can use Data Protector again in the RMAN mode, you have to re-link the
Oracle again with the Data Protector Database Library.
b. You have to define the parameter init in the initialization file initORACLE_SID .
ora .
Execute the following commands: brrestore -d pipe -u user/password -t online -m all brrestore -d disk -u user/password
If this fails, see the SAP Online Help to learn how to execute backup and restore directly to disk using the SAP backup utility. Check the error message and resolve this issues before you continue.
4. Verify that the SAP backup tools correctly start backint (which is provided by Data
Protector):
Move the original backint and create a test script named backint in the directory where the SAP backup utility resides, with the following entries:
#!/usr/bin/sh echo "Test backint called as follows:" echo "$0 $*" echo "exiting 3 for a failure" exit 3
Data Protector (10.00) Page 178 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
: brbackup -t offline -d util_file -u user/password -c
If you receive backint arguments, this means that SAP is properly configured for backup using backint; otherwise you have to reconfigure SAP.
See Configuring SAP R/3 databases, on page 139
.
Configuration problems
IMPORTANT:
The procedure described in the previous sections must be performed before you start checking the Data Protector configuration.
1. Verify that the Data Protector software has been installed properly.
For details, see the HPE Data Protector Installation Guide.
2. Perform a filesystem backup of the SAP R/3 Database Server:
Perform a filesystem backup of the SAP Database Server system so that you can eliminate any potential communication problems between the SAP Database Server and the Data Protector Cell
Manager system.
Do not start troubleshooting an online database backup unless you have successfully completed a filesystem backup of the SAP Database Server system.
See the HPE Data Protector Help index “standard backup procedure” for details about how to do a filesystem backup.
3. Examine the environment variables:
If you need to export some variables before starting the Oracle Server Manager, TNS listener, or other Oracle utility, these variables must be defined in the Environment section of the Data
4. Verify the permissions of the currently used user account:
Your user account has to enable you to perform backup or restore using Data Protector. Use the testbar2 utility to check the permissions:
/opt/omni/bin/utilns/testbar2 -perform:checkuser
If the user account holds all required permissions, you will receive only NORMAL messages displayed on the screen.
See also Configuring user accounts, on page 133 .
5. Examine system errors:
System errors are reported in the
/var/opt/omni/log/debug.log
(HP-UX, Solaris, and Linux systems) or /usr/omni/log/debug.log
(other UNIX systems) file on the SAP Server.
Problem
Configuration fails due to a script failure
During configuration of an SAP R/3 database, Data Protector reports the following error:
Integration cannot be configured.
Data Protector (10.00) Page 179 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Script failed. Cannot get information from remote host.
Action
Backup problems
At this stage, you should have performed all the verification steps described in the previous sections. If backup still fails, proceed as follows:
1. Check your SAP Server configuration:
To check the configuration, start the following command on the SAP Server system:
/opt/omni/lbin/util_sap.exe -CHKCONF ORACLE_SID
(HP-UX, Solaris, and Linux systems) or
/usr/omni/bin/util_sap.exe -CHKCONF ORACLE_SID (other UNIX systems)
In case of an error, the error number is displayed in the form *RETVAL*Error_number .
To get the error description, start the command:
/opt/omni/lbin/omnigetmsg 12 Error_number
(HP-UX, Solaris, and Linux systems) or
/usr/omni/bin/omnigetmsg 12 Error_number (other UNIX systems)
The message *RETVAL*0 indicates successful configuration.
2. Verify Data Protector internal data transfer using the testbar2 utility.
Before you run the testbar2
utility, verify that the Cell Manager name is correctly defined on the
SAP Database Server. Check the /etc/opt/omni/client/cell_server (HP-UX, Solaris, and
Linux systems) or /usr/omni/config/cell/cell_server (other UNIX systems) file, which contains the name of the Cell Manager system. Then execute the following command:
/opt/omni/bin/utilns/testbar2 -type:SAP -appname:ORACLE_SID -bar:backup_
specification_name -perform:backup (HP-UX, Solaris, and Linux systems)
/usr/omni/bin/utilns/testbar2 -type:SAP -appname:ORACLE_SID -bar:backup_
specification_name -perform:backup (other UNIX systems)
Examine the errors reported by the testbar2
utility by clicking the Details button in the Data
Protector Monitor context.
If the messages indicate problems concerning the Data Protector side of the integration, proceed as follows:
a. Check that the owner of the backup specification is the Oracle OS user described in
Configuring user accounts, on page 133
b. Check that the respective Data Protector user group has the See private objects user right enabled.
c. Create an SAP backup specification to back up to a null or file device. If the backup succeeds, the problem may be related to the backup devices.
For instructions on troubleshooting devices, see the HPE Data Protector Troubleshooting
Guide.
If the test fails again, call support.
3. Verify the backup using backint
Data Protector (10.00) Page 180 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration export OB2BARLIST=barlist_name export OB2APPNAME=ORACLE_SID
/opt/omni/lbin/backint -f backup -t file -u ORACLE_SID -i input_file (HP-UX,
Solaris, and Linux systems)
/usr/omni/bin/backint -f backup -t file -u ORACLE_SID -i input_file (other UNIX systems) where input_file is a file with a list of full pathnames for backup.
Backint expects the list of files in the following format:pathName_1pathName_2pathName_3
Problem
Util_File_Online SAP backup fails with “semop() error”
When the util_file_online option is used with BRBACKUP (for example, if you select the
Brbackup_Util_File_Online template), the tablespaces are switched into/from backup mode individually. As there can be only one process communicating with BRBACKUP, several sapback processes are using a semaphore to synchronize their interaction with BRBACKUP.
The number of sapback processes is calculated as the sum of concurrencies of all devices used for backup. With a large number of sapback processes, the maximum number of processes that can have undo operations pending on any given IPC semaphore on the system may be exceeded. In such case, several sapback agents will fail with the following error:
[28] No space left on device.
Action
Perform any of the following actions to resolve the problem: l
Reduce the number of backup devices or their concurrency.
l
Increase the value of the semmnu kernel parameter. After you increase the value, rebuild the kernel and restart the system.
Problem
Backup fails with “Connect to database instance failed”
If you start a backup while the database instance is in the unmount or mount mode, the session fails with a message similar to the following:
BR0301E SQL error -1033 at location BrDbConnect-2
ORA-01033: ORACLE initialization or shutdown in progress
BR0310E Connect to database instance HOOHOO failed
Action
Before you start a backup, ensure that the database instance is in the open or shutdown mode.
Restore problems
At this stage, you should have performed all the verification steps described in the previous sections.
After this, proceed as follows:
Data Protector (10.00) Page 181 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
1. Verify a user for the restore:
Verify that user specified for the restore session is the user of backup session and that he/she belongs to the Data Protector operator or admin group.
See Configuring user accounts, on page 133
.
2. Verify that a backup object exists on the backup media and in the IDB:
This can be done by executing the command omnidb -sap "object_name" -session "Session_ID" -media (HP-UX, Solaris, and Linux systems) or omnidb -sap "object_name" -session "Session_ID" -media (other UNIX systems) on the SAP Database Server system.
The output of the command lists detailed information about the specified backup object, session
IDs of the backup sessions containing this object, and a list of the media used.
For detailed syntax of the omnidb
command, execute: omnidb -help (HP-UX, Solaris, and Linux systems) omnidb -help (other UNIX systems)
You can also do this using the SAP tools:
Use backint, so that SAP tools will also use this command to query:
/opt/omni/lbin/backint -f inquiry -u ORACLE_SID -i input_file (HP-UX, Solaris, and
Linux systems)
/usr/omni/bin/backint -f inquiry -u ORACLE_SID -i input_file (other UNIX systems) where the specified
input_file
is queried.
If this fails, check if the backup session was performed successfully and if the query was started under the appropriate user account.
Backint anticipates a list of files of the following format:
backup_ID_1 pathName_1
[
targetDirectory_1
]
backup_ID_2 pathName_2
[
targetDirectory_2
]
backup_ID_3 pathName_3
[
targetDirectory_3
]
To retrieve the
backup_ID
numbers, enter the following command: echo "#NULL #NULL" | backint –f inquiry –u ORACLE_SID or, alternatively, you can just specify #NULL as
backup_ID_1
in the
input_file
. In this case, the latest backup session for the file is used for the restore.
3. Verify the restore using the Data Protector user interface
This test is possible if the objects have been backed up by backint.
If this fails, check if the backup session was performed successfully and if the query was started under the appropriate user account.
4. Simulate a restore session
Once you know the information about the object to be restored, you can simulate a restore using the Data Protector testbar2 utility.
Before you run testbar2 , verify that the Cell Manager name is correctly defined on the SAP
Database Server.
Data Protector (10.00) Page 182 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Check the /etc/opt/omni/client/cell_server (HP-UX, Solaris, and Linux systems) or
/usr/omni/config/cell/cell_server
(other UNIX systems) file, which contains the name of the Cell Manager system.
Then, test the Data Protector internal data transfer using the testbar2 utility:
/opt/omni/bin/utilns/testbar2 -type:SAP
-appname:ORACLE_SID
-perform:restore
-object:object_name
-version:object_version
-bar:backup_specification_name (HP-UX, Solaris, and Linux systems) or
/usr/omni/bin/utilns/testbar2 -type:SAP
-appname:ORACLE_SID
-perform:restore
-object:object_name
-version:object_version
-bar:backup_specification_name (other UNIX systems)
You should see only NORMAL messages displayed on your screen, otherwise examine the errors reported by the testbar2 utility by clicking the Details button in the Data Protector Monitor context.
5. Verify the restore using backint
Execute the following command:
HP-UX, Solaris, and Linux systems:
/opt/omni/lbin/backint -f restore -u ORACLE_SID
-i input_file
Other UNIX systems:
/usr/omni/bin/backint -f restore -u ORACLE_SID -i input_file where the contents of the
input_file
will be restored.
If this fails, check if the session was performed successfully and if the restore was started under the appropriate user account.
Backint anticipates a list of files in the following format:
backup_ID_1
pathName_1
[targetDirectory_1]backup_ID_2pathName_2 [targetDirectory_2]backup_ID_3pathName_3
[targetDirectory_3]
To retrieve the
backup_ID
numbers, enter the following command: echo #NULL #NULL | backint –f inquiry –u ORACLE_SID
Problem
Restore of SAP R/3 tablespaces located on raw partitions fails
When restoring SAP tablespaces that are located on raw partitions using the Data Protector GUI, the restore fails with a message similar to the following:
[Major] From: [email protected] "SAP" Time: 5/9/06 3:33:51 PM
/dev/sapdata/rsapdata Cannot restore -> rawdisk section !
[Warning] From: [email protected] "SAP"
Time: 5/9/06 3:42:45 PM Nothing restored.
Data Protector (10.00) Page 183 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Action
Use SAP commands (for example, brrestore
) to restore these tablespaces.
Verifying the prerequisites (Oracle side)
Perform the following verification steps, in numerical order, to verify that Oracle is installed properly:
1. On the application system, verify that the target database is online, as follows:
UNIX systems:
export ORACLE_SID=Oracle_SID export ORACLE_HOME=Oracle_home_path
$ORACLE_HOME/bin/sqlplus
Windows systems:
set ORACLE_SID=Oracle_SID set ORACLE_HOME=Oracle_home_path
%ORACLE_HOME%\bin\sqlplus
At the SQLPlus prompt, type: connect user/passwd@service select * from dba_tablespaces exit;
Try starting the target database.
2. In order to establish the TNS network connection, verify that Net8 software is configured correctly for the target database, as follows: l
On the application system, perform the following:
UNIX systems:
$ORACLE_HOME/bin/lsnrctl status service
Windows systems:
%ORACLE_HOME%\bin\lsnrctl status service
If it fails, either start the TNS listener process or see the Oracle documentation on how to create the TNS configuration file (LISTENER.ORA).
l
On the application system, perform the following. Use sqlplus :
UNIX systems:
export ORACLE_SID=Oracle_SID export ORACLE_HOME=Oracle_home_path
$ORACLE_HOME/bin/sqlplus
Windows systems:
set ORACLE_SID=Oracle_SID set ORACLE_HOME=Oracle_home_path
Data Protector (10.00) Page 184 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
%ORACLE_HOME%\bin\sqlplus
At the SQLPlus prompt, type: connect user/passwd@service ; exit;
If it fails, see the Oracle documentation on how to create the TNS configuration file
(TSNAMES.ORA).
Verifying the prerequisites (SAP side)
.
Perform the following verification steps, in numerical order, to verify that SAP is installed properly:
1. On the application system, verify a backup directly to disk, as follows: brbackup -d disk -u user/password
If it fails, see the SAP Online Help for instructions on how to execute a backup to disk using the
SAP backup utility.
2. On the application system, verify a restore from the disk, as follows: brrestore -d disk -u user/password
If it fails, see the SAP Online Help for instructions on how to execute a restore to disk using the
SAP restore utility.
3. On the application system, verify that SAP is configured properly, as follows:
Move the original backint . Create a test script with the name backint in the directory with the
SAP backup utility, with the following entries:
#!/usr/bin/sh echo "Test backint called as follows:" echo "$0 $*" echo "exiting 3 for a failure" exit 3
Export all environment variables required by the SAP (SAPDATA_HOME, SAPBACKUP...) and then start the command with the backup owner user: brbackup -t offline_split -d util_file -u user/password -c or, if Data Protector uses splitint : brbackup -t offline_mirror -d util_file -u user/password -c
If you receive arguments from backint
, that means SAP is properly configured for backup using backint . Otherwise, you should reconfigure SAP.
Verifying the configuration
Before you begin this section, be sure that you completed all the steps provided in the sections
Verifying the prerequisites (Oracle side) , on the previous page
and Verifying the prerequisites (SAP side) , above .
Data Protector (10.00) Page 185 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Perform the following verification steps, in numerical order, to verify that Data Protector is configured properly:
1. On the application system, verify a Data Protector filesystem backup of the SAP Database
Server:
Perform a filesystem backup of the Oracle Server system so that you can eliminate any potential communication problems between the Oracle Server and the Data Protector Cell Manager system.
See the HPE Data Protector Help index “standard backup procedure” for details about how to do a filesystem backup.
If it fails, see the HPE Data Protector Troubleshooting Guide for help with troubleshooting a filesystem backup.
2. Verify the environment variable on the application system:
If you have to export some variables before starting the SAP backup utilities, Oracle Server
Manager, or the TNS listener, set these environment variables using the Data Protector GUI.
3. Verify the permissions of the SAP user on application system:
SAP user permissions must be set to enable you to perform an SAP backup or restore with Data
Protector. See Configuring user accounts, on page 133 . Use the
testbar2 to check the permissions: l
Login in as an SAP user l
Execute /opt/omni/bin/testbar2 -perform:checkuser
If the user account has all the required permissions, you will see only the usual messages displayed on the screen.
4. Examine the system errors:
System errors are reported in the following file on the Oracle Server:
/var/opt/omni/log/debug.log
Verifying the backup configuration
Before you begin this section, be sure that you completed all the steps provided in the sections
Verifying the prerequisites (Oracle side) , on page 184
and Verifying the prerequisites (SAP side) , on the previous page
.
Perform the following verification steps, in numerical order, to verify that Data Protector is configured properly:
1. Verify the Data Protector SAP ZDB configuration on the application system:
Execute the following command:
HP-UX and Solaris systems : /opt/omni/lbin/util_sap –CHKCONF ORACLE_SID
Windows systems: Data_Protector_home\bin\util_sap.exe –CHKCONF ORACLE_SID
If an error occurs, the error number is displayed in the form *RETVAL*error_number .
To get the error description, on the Cell Manager, execute:
Windows systems:
Data_Protector_home\bin\omnigetmsg 12 error_number
Data Protector (10.00) Page 186 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration which is located on the Cell Manager.
HP-UX and Linux systems: /opt/omni/lbin/omnigetmsg 12 error_number
2. Verify the SAP user.
Check that the respective user group has the
See Private Objects
user right selected. See also
Configuring user accounts, on page 133
.
3. On the application system, verify the backup using testbar2
:
Execute the following to ensure that communication within Data Protector is established: l
Create a non-ZDB backup specification on the application system.
l
Execute:
/opt/omni/bin/testbar2 –type:SAP –appname:ORACLE_SID \ –perform:backup – file:file_name -bar barlist_name
If it fails, check the errors and try to fix them or call a support representative for assistance.
4. On the application system, verify the backup using backint:
Execute the following command to ensure that communication within Data Protector is established and that a backup of files can be performed: l
Create a non-ZDB backup specification on the backup system.
l export OB2BARLIST=barlist_name export OB2APPNAME=ORACLE_SID
/opt/omni/lbin/backint –f backup –t file –u ORACLE_SID -i \ input_file where
input_file
is the file containing the full pathnames for backup.
If it fails, check the errors and try to fix them or call a support representative for assistance.
Verifying restore
Before you begin this section, be sure that you completed all the steps provided in the sections
Verifying the prerequisites (Oracle side) , on page 184
and Verifying the prerequisites (SAP side) , on page 185 .
Perform the following verification steps, in numerical order, to verify that Data Protector is configured properly:
1. Verify the user for the restore
Verify that the user specified for the restore session is the user of the backup session and that they belong to the Data Protector operator or admin group. Check that the respective user group has the See private objects user right selected.
2. Verify that files are backed up and in the Data Protector database: l
Using the omnidb
command;
See the appropriate man page on using the omnidb command.
l
Using backint ;
Data Protector (10.00) Page 187 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
SAP tools also use this command to make a query.
/opt/omni/lbin/backint –f inquiry –u ORACLE_SID -i input_file where
input_file
is what will be queried. Backint expects a list of files in the following format:
backint_ID_1 pathName_1
backint_ID_2 pathName_2
backint_ID_3 pathName_3
To retrieve the backint_ID numbers, enter the following command: echo "#NULL #NULL" | backint –f inquiry –u ORACLE_SID or, alternatively, you can just specify #NULL as backint_ID_1 in the input_file. In this case, the latest backup session for the file is used for the restore.
If it fails, proceed as follows: l
Check the backup session - was it successful?
l
Check the user rights. Was the query started under the correct SAP user account?
l
Call a support representative for assistance.
3. Verify the restore using Data Protector or CLI:
If it fails, proceed as follows: l
Check the backup session - was it successful?
l
Check that the files are in the Data Protector database.
l
Check the user rights. Was the restore started under the correct SAP user account?
l
Call a support representative for assistance.
4. Verify the restore using testbar2
:
Execute the following to ensure that restore is possible:
/opt/omni/bin/testbar2 –type:SAP –appname:ORACLE_SID \ –perform:restore – file:file_name -bar barlist_name -object objectName
If it fails, proceed as follows: l
Check the backup session - was it successful?
l
Check that the files are in the Data Protector database.
l
Check the user rights. Was the a restore started under the correct SAP user account l
Call a support representative for assistance.
5. Verify the restore using backint : backint
is the same command used by the SAP backup utility.
/opt/omni/lbin/backint –f restore –u ORACLE_SID -i input_file
Data Protector (10.00) Page 188 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration where input_file specifies what will be restored; backint expects the list of files in the following format:
backint_ID_1
pathName_1 [targetDirectory_1]
backint_ID_2pathName_2 [targetDirectory_2]
backint_ID_3pathName_3 [targetDirectory_3]
To retrieve the
backint_ID
numbers, enter the following command: echo "#NULL #NULL" | backint –f inquiry –u ORACLE_SID or, alternatively, you can just specify #NULL as
backint_ID_1
in the
input_file
. In this case, the latest backup session for the file is used for the restore.
If it fails, proceed as follows: l
Check the backup session - was it successful?
l
Check that the files are in the Data Protector database.
l
Check the user rights. Was the restore started under the correct SAP user account?
l
Call a support representative for assistance.
Configuration and backup problems
The following list gives a description of problems and actions to be taken to resolve them: l
The Server Manager is unable to connect to the destination
Check whether the Oracle TNS listener process is up and running. Check whether there are any environment variables required for a successful remote connection to the target database; for example,
TNS_ADMIN
and SHLIB_PATH. Set these environment variables using the Data Protector
GUI.
.
l
Configuration procedure fails
Check whether the Oracle Server is up and running.
Check the login information for the target from the application system using Oracle Server Manager.
If you cannot log in, then perform the following actions:
Check whether sysoper and sysdba rights are set for the Oracle administrator user.
Examine system errors reported in the debug.log
, sap.log
and oracle8.log
files, located in the default Data Protector log files directory.
If you have special Oracle environment settings, ensure that they are registered in the Environment sublist of the Data Protector SAP configuration file:
/etc/opt/omni/server/integ/config/SAP/
client_name
ORACLE_SID (UNIX Cell Manager), or Data_Protector_program_data \Config\server\integ\config\ sap\
client_name
%
ORACLE_
SID (Windows Cell Manager).
.
Data Protector (10.00) Page 189 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration l
Starting the backup fails
On UNIX systems, check the output of the following command on the application system:
/opt/omni/lbin/util_sap.exe -CHKCONF ORACLE_SID
In case of an error, the error number is displayed in the form:
*RETVAL*Error_number
To get the error description, start the following command on the application system:
/opt/omni/lbin/omnigetmsg 12 Error_number
On Windows systems, perform the following procedure using the Data Protector GUI:
1. In the Context List, select Backup.
2. In the Scoping Pane, expand Backup, Backup Specifications, and then SAP R/3. A list of
SAP backup specifications is displayed.
3. In the Scoping Pane, select the failed backup specification and right-click on the SAP R/3 server item in the Results Pane to display a pop-up menu.
4. From the pop-up menu, select Check Configuration.
A short description of the problems and how to resolve them is displayed.
l
Backup does not work
o
Check whether the Cell Manager is correctly set on the application system. The file
/etc/opt/omni/client/cell_server (UNIX systems) or HKEY_LOCAL_
MACHINE\SOFTWARE\Hewlett-Packard Enterprise\OpenView\OmniBackII\
Site\CellServer
(Windows systems) must contain the name of the Cell Manager.
o
Check that the primary_db
parameter in the initORACLE_SID.sap
file on the application system is set to
LOCAL
.
o
On UNIX systems, check whether the users are properly configured in user groups. Both the
UNIX Oracle administrator ( ora ORACLE_SID) and UNIX SAP administrator (ORACLE_
SID
adm ) have to be in the Data Protector operator class.
o
On UNIX systems, check whether the permissions of the SAPDATA_HOME /sapbackup/ directory are set to 755.
o
On Windows systems, check that the user account that started the Data Protector Inet service is added in the Data Protector operator class.
l
Backup fails with “Connect to database instance failed”
If you start a backup while the database instance is in the unmount or mount mode, the session fails with a message similar to the following:
BR0301E SQL error -1033 at location BrDbConnect-2
ORA-01033: ORACLE initialization or shutdown in progress
BR0310E Connect to database instance HOOHOO failed
Before you start a backup, ensure that the database instance is the open or shutdown mode.
Problem
Configuration fails due to a script failure
Data Protector (10.00) Page 190 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
During configuration of an SAP R/3 database, Data Protector reports the following error:
Integration cannot be configured.
Script failed. Cannot get information from remote host.
Action
If the SAP R/3 database is located on a Windows system, check the environment settings and ensure
Data Protector Inet is running under a user account which has the required privileges. For details, see
Before you begin, on page 133 .
If the SAP R/3 database is located on a UNIX system, resolve the problem by reviewing the user
account configuration. For details, see Configuring user accounts, on page 133
.
Problem
Backup using backint fails on Solaris and HP-UX
A backup using backint fails on a Solaris and HP-UX systems with the following error:
[Major] From: [email protected] "SAP" Time: 4/29/09 3:55:52 PM
Cannot open file '/saphome/SAP/sapbackup/cntrlSAP.dbf'. Error: 13
Action
Share the directories on the application system through NFS with root permissions.
On the application system, add the following line in the file /etc/dfs/dfstab :
Solaris systems:
share -F nfs -o anon=0 /usr/src
HP-UX systems:
share -F nfs -o anon=0,rw -d "" SAPHOME
On the application system, add the following line in the file
/etc/exports
:
RHEL/SUSE systems:
Problem
A ZDB session fails after reporting connections errors
A ZDB session reports the following critical errors and completes with failures:
Connection with DMA was reset
Signal SIGABRT (6) received from BRtools
Action
In the SAP parameter file on the application system, set the parameter primary_db to LOCAL and restart the session.
For details, see Configuring the SAP R/3 parameter file , on page 146
.
Data Protector (10.00) Page 191 of 413
Zero Downtime Backup Integration Guide
Chapter 2: Data Protector SAP R/3 ZDB integration
Restore problems
Problem
ZDB, restore, or instant recovery sessions fail due to invalid characters in filenames
On Windows systems, where the Oracle Database Character Set (DBCS) is not set to the same value as the default Windows character set for non-Unicode programs, and where SAP tools are used to create Oracle datafiles, ZDB, restore, and instant recovery fail if the datafiles contain non-ASCII or non-Latin 1 characters.
Actions
Use any of the following solutions: l
For new Oracle installations, set the DBCS to UTF-8.
l
If you do not use other non-Unicode programs, set the language for non-Unicode programs to the same value as DBCS.
l
Do not use non-ASCII or non-Latin 1 characters for filenames.
Problem
Restore of SAP R/3 tablespaces located on raw partitions fails
When restoring SAP tablespaces that are located on raw partitions using the Data Protector GUI, the restore fails with a message similar to the following:
[Major] From: [email protected] "SAP" Time: 5/9/06 3:33:51 PM
/dev/sapdata/rsapdata Cannot restore -> rawdisk section ![Warning] From:
[email protected] "SAP" Time: 5/9/06 3:42:45 PM Nothing restored.
Action
Use SAP commands (for example, brrestore ) to restore these tablespaces.
Data Protector (10.00) Page 192 of 413
Chapter 3: Data Protector Microsoft SQL
Server ZDB integration
Introduction
This chapter explains how to configure and use the Data Protector Microsoft SQL Server ZDB integration. It describes the concepts and methods you need to understand to back up and restore the Microsoft SQL
Server (SQL Server) database objects.
During the backup, an SQL Server snapshot is made (the database files are frozen and any transactions to them are cached), so the database is highly available (online backup). The I/O to it is suspended during the time it takes to create a replica (split the mirror disks or create snapshots).
NOTE:
SQL Server snapshot is an SQL Server related term and does not mean the same as a disk array snapshot.
The following disk arrays and array configurations are supported:
Supported array
EMC Symmetrix (EMC)
HPE P9000 XP Disk Array Family (P9000 XP Array)
HPE P6000 EVA Disk Array Family (P6000 EVA Array)
Supported configurations
Dual-host TimeFinder
HPE BC P9000 XP, HPE CA P9000 XP, combined HPE CA+BC P9000 XP
HPE BC P6000 EVA, combined HPE
CA+BC P6000 EVA
Non–HPE Storage Arrays (NetApp, EMC VNX, EMC VMAX) Local replication
All ZDB types (ZDB to tape, ZDB to disk, and ZDB to disk+tape) are supported by this integration. For ZDB types description, see the HPE Data Protector Concepts Guide and the HPE Data Protector Help.
Using Data Protector, you can restore your SQL Server data: l
From backup media to the application system on LAN (standard restore).
l
Using the instant recovery functionality.
The following table gives an overview of SQL Server recovery methods:
ZDB type
ZDB to tape
ZDB to disk
ZDB to disk+tape
Recovery method
Standard restore
Instant recovery
Standard restore, instant recovery
For a description of ZDB and instant recovery (IR) concepts, see the HPE Data Protector Concepts Guide.
ZDB and IR do not support an SQL Server availability group configuration.
Data Protector (10.00) Page 193 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Integration concepts
Data Protector integrates with SQL Server through the Data Protector sql_bar.exe
executable, installed on SQL Server. During backup, sql_bar.exe
, started on the application system, connects to
SQL Server to find the locations of the database files. The integration then backs up SQL Server database(s), which are replicated within a disk array.
During restore, sql_bar.exe
connects to SQL Server to receive the restore data, which is then written to disks.
ZDB process depends on whether you replicate your data in a split mirror or snapshot replication and on the selected ZDB type. Restore process depends on the restore type - standard restore or instant recovery. See the HPE Data Protector Concepts Guide for a detailed description of replication techniques and ZDB and restore processes.
Backup and restore concepts
Data Protector (10.00) Page 194 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Configuring the integration
Prerequisites
l
You need a license to use the SQL Server ZDB integration. For information, see the HPE Data
Protector Installation Guide.
l
Make sure that you correctly installed and configured SQL Server.
o
For supported versions, platforms, devices, and other information, see the latest support matrices at https://softwaresupport.hpe.com/manuals .
o
For information on installing, configuring, and using SQL Server, see the SQL Server documentation.
l
Make sure that you correctly installed Data Protector. For information on installing Data Protector in various architectures and installing a Data Protector disk array integration (P6000 EVA Array, P9000
XP Array, EMC, or NetApp Storage) with SQL Server, see the HPE Data Protector Installation
Guide.
Every SQL Server to be used with Data Protector must have the MS SQL Integration component installed.
l
Install SQL Server on the application system. Install user databases on the disk array source volumes (system databases can be installed anywhere). If the system databases are also installed on a disk array, they must reside on different source volumes than user databases.
If SQL Server is installed on the backup system as well, its databases must reside on source volumes different from the source volumes used for this integration. Drive letters/mount points assigned to those volumes must also be different from the drive letters/mount points assigned to the volumes used for this integration.
Before you begin
l
Configure devices and media for use with Data Protector. For instructions, see the HPE Data
Protector Help index: “configuring devices” and “creating media pools”. l
If you plan to use Integrated authentication to connect to an SQL Server instance, you need to restart the Data Protector Inet service under a Windows domain user account that has the appropriate SQL Server permissions for running backup and restore sessions. For supported
Windows operating systems, use user impersonation. For details on setting accounts for the Inet service user impersonation, see the HPE Data Protector Help index: “Inet user impersonation”.
l
Using the SQL Server Management Studio, add the user account which you will use for backing up and restoring SQL Server data to the fixed server role sysadmin
. For instructions, see the SQL
Server documentation.
l
To test whether SQL Server and Cell Manager communicate properly, configure and run a Data
Protector filesystem ZDB and restore. For instructions, see the HPE Data Protector Help.
Data Protector (10.00) Page 195 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Data Protector SQL Server configuration file
Data Protector stores integration parameters for every configured SQL Server on the Cell Manager in:
HP-UX and Linux systems:
/etc/opt/omni/server/integ/config/MSSQL/ClientName%InstanceName
Windows systems:
Data_Protector_program_data
\Config\Server\Integ\Config\MSSQL\ClientName%InstanceName
Configuration parameters are the username and password of the SQL Server user, who must have permissions to run backups and restores within SQL Server (assuming the standard security is used).
They are written to the Data Protector SQL Server configuration file during configuration of the integration.
The content of the configuration file is:
Login=' user';
Password=' encoded_password';
Domain='
domain';
Port='
PortNumber';
IMPORTANT:
To avoid backup problems, make sure that the syntax of your configuration file matches the examples.
Examples l
SQL Server authentication:
Login='sa';
Domain='';
Password='jsk74yh80fh43kdf'; l
Windows authentication:
Login='Administrator';
Domain='IPR';
Password='dsjf08m80fh43kdf'; l
Integrated authentication:
Login='';
Domain='';
Password='kf8u3hdgtfh43kdf';
Configuring an SQL Server cluster
In a cluster, all the nodes must be installed as Data Protector cluster-aware clients and the Data
Protector Inet
service on all nodes must run under a Windows domain user account that has also cluster administrator rights.
Data Protector (10.00) Page 196 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
You must configure the
Data Protector Inet
service user impersonation for all cluster nodes. The
Windows domain user account that is used must be given the following Windows operating system
Security Policy privileges: l
Impersonate a client after authentication l
Replace a process level token
For more information, see the HPE Data Protector Help index: “cluster-aware client”, “Inet user impersonation”, and the SQL Server cluster documentation.
Configuring SQL Server instances
An SQL Server instance is configured during the creation of the first backup specification. The configuration consists of setting the user account that Data Protector should use to connect to the SQL
Server instance. The specified login information is saved to the Data Protector SQL Server instance configuration file on the Cell Manager.
NOTE:
Make sure that the user account to be used has appropriate SQL Server permissions for running backups and restores. Check the permissions using SQL Server Enterprise Manager.
.
Prerequisites l
SQL Server must be online during configuration.
l
Make sure that the SQL Server Browser service is running.
l
Configuration must be performed for every SQL Server instance separately.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click MS SQL Server, and click Add
Backup.
3. In the Create New Backup dialog box, select the Blank Microsoft SQL Server Backup
.
Click OK.
.
5. In Application database, select or specify the name of the SQL Server instance.
Windows Server 2008: If you intend to use Integrated authentication and you want that the backup session to run under the specified operating system user account, specify the Specify OS
user option. For information on the User and group/domain options, press F1.
Click Next.
6. In the Configure MS SQL Server dialog box, specify the user account that Data Protector should
Data Protector (10.00) Page 197 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration use to connect to the SQL Server instance.
l
SQL Server authentication: SQL Server user account. Specify a username and password.
l
Windows authentication: Windows domain user account (preferred option). Specify a username, password, and the domain.
l
Integrated authentication: Select this option to enable Data Protector to connect to the SQL
Server instance with the following Windows domain user account: o
Windows Server 2008: The account specified in the User and group/domain options in the previous step or in the Client selection page.
o
Other Windows systems: The account under which the Data Protector Inet service on the SQL Server system is running.
Make sure that the user account you specify has the appropriate permissions for backing up and restoring the SQL Server databases.
See Configuring SQL Server, below .
Configuring SQL Server
NOTE:
It is recommended that the SQL Server system administrator configures the integration.
Data Protector (10.00) Page 198 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
For details about security, see the SQL Server documentation.
Click OK to confirm the configuration.
7. The SQL Server instance is configured. Exit the GUI or proceed with creating the backup
specification at Creating ZDB specifications, on page 201 .
Using the Data Protector CLI
Execute: sql_bar config [-appsrv:SQLServerClient] [-instance:InstanceName] [dbuser:SQLServerUser -password:password | -dbuser:WindowsUser -password:password domain:domain]
Parameter description
-appsrv:SQLServerClient
-instance:InstanceName
The client system on which the SQL Server instance is running.
This option is not required if you execute the command locally.
The SQL Server instance name. If you omit this option, the default SQL Server instance is configured.
The SQL Server user account (SQL Server authentication) -dbuser:SQLServerUser password:password
-dbuser:WindowsUser password:password domain:domain
The Windows domain user account (Windows authentication)
NOTE:
If no user account is specified, Data Protector uses Integrated authentication.
The message *RETVAL*0 indicates successful configuration.
Changing and checking configuration
You can check and change configuration using the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications and then MS SQL Server. Click a backup specification for which you want to change the configuration.
3. In the Source property page, right-click the SQL Server name and select Configure.
4. Configure SQL Server as described in Configuring SQL Server instances, on page 197
.
5. Right-click SQL Server and select Check Configuration. See Checking configuration, below
.
Checking configuration
Data Protector (10.00) Page 199 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Using the Data Protector CLI
To change the configuration, execute the command for configuring SQL Server instances again, entering different data.
To check configuration, execute: sql_bar chkconf [-instance:InstanceName]
If the optional parameter -instance:InstanceName is not specified, the default instance is checked.
If the integration is not properly configured, the command returns:
*RETVAL*8523
To get the information about the existing configuration , execute: sql_bar getconf [-instance:InstanceName]
If -instance:InstanceName is not specified, Data Protector returns configuration for the default instance.
Backup
To run ZDB of an existing SQL Server ZDB specification:
Data Protector (10.00) Page 200 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration l
Schedule a backup using the Data Protector Scheduler.
l
Start an interactive backup using the Data Protector GUI or CLI.
Prerequisites
l
In case of nested mount points, the same drive letters, on which the source volumes to be replicated reside on the application system, must exist on the backup system to enable successful mounting of the target volumes. If the same drive letters do not exist on the backup system, the backup fails.
Considerations
Your session will fail if: l
You start ZDB, restore, or instant recovery using the same source volume on the application system at the same time. A session must be started only after the preceding session using the same source volume on the application system finishes.
l
SQL Server services are not running when the backup starts.
To configure a ZDB, create a Data Protector SQL Server ZDB specification.
Creating ZDB specifications
Create a ZDB specification, using the Data Protector Manager.
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click MS SQL Server, and click Add
Backup.
3. In the Create New Backup dialog box, select the Blank Microsoft SQL Server Backup template.
From the Backup type drop-down list, select Snapshot or split mirror backup, and from the Sub
type drop-down list, select the appropriate disk array agent. The agent must be installed on the
Data Protector (10.00) Page 201 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Selecting a Microsoft SQL Server backup template and the snapshot or split mirror backup
Click OK.
4. Under Client systems, select the SQL Server system. In cluster environments, select the virtual server of the SQL Server resource group.
In the Backup system drop-down list, select the backup system.
for P9000 XP Array, P6000 EVA Array backup options, on page 205 for P6000 EVA
Array, NetApp Storage backup options, on page 207
for NetApp Storage, EMC VNX Storage backup options, on page 208
for EMC VNX Storage, or EMC VMAX Storage backup options, on page 209
for EMC VMAX Storage). For detailed information on the backup options, press F1.
EMC:
In the EMC GeoSpan for Microsoft Cluster Service environment, select the backup system for the active node and specify the TimeFinder configuration.
After a failover in EMC GeoSpan for MSCS, select the backup system for the currently active node and save the backup specification.
Data Protector (10.00) Page 202 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
EMC backup options
P9000 XP Array:
To enable instant recovery, leave Track the replica for instant recovery selected.
Data Protector (10.00) Page 203 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
P9000 XP Array backup options
P6000 EVA Array:
To enable instant recovery, select Track the replica for instant recovery.
Data Protector (10.00) Page 204 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
P6000 EVA Array backup options
Data Protector (10.00) Page 205 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
P10000 3PAR Array specifics:
P10000 3PAR Array backup options
Data Protector (10.00) Page 206 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
NetApp Storage:
NetApp Storage backup options
Data Protector (10.00) Page 207 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
EMC VNX Storage:
EMC VNX Storage backup options
Data Protector (10.00) Page 208 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
EMC VMAX Storage:
EMC VMAX Storage backup options
Click Next.
5. In Application database, specify the name of the SQL Server instance.
Windows Server 2008: If you intend to use Integrated authentication and you want that the backup session runs under the specified operating system user account, specify the Specify OS
user option. For information on the User and group/domain options, press F1.
Click Next.
6. If the client is not configured, the Configure MS SQL Server dialog box appears. Configure it as
described in Configuring SQL Server instances, on page 197 .
7. Select the databases to be backed up.
IMPORTANT:
To enable instant recovery, create different backup specifications for user and system databases.
Data Protector (10.00) Page 209 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Selecting user databases
Click Next.
8. Select the devices. Click Properties to set the media pool and preallocation policy. The device concurrency is set to 1 and cannot be changed. For more information on options, press F1.
To create additional backup copies (mirrors), specify the desired number by clicking Add
mirror/Remove mirror. Select separate devices for each mirror. The minimum number of devices for mirroring equals the number of devices used for backup.
For more information on object mirroring, see the HPE Data Protector Help.
NOTE:
Object mirroring is not supported for ZDB to disk.
Click Next.
9. Select backup options.
For information on Backup Specification Options and Common Application Options, see the
HPE Data Protector Help.
Click Next.
10. Click Save As to save the backup specification, specify a name and a backup specification group. Optionally, you can click Save and Schedule to save, and then schedule the backup specification. For more information on how to create and edit schedules, see Scheduler in Data
Data Protector (10.00) Page 210 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Protector in HPE Data Protector Administrator's Guide.
Note that only Full backup is performed.
You start the backup specification by clicking Start Backup.
Saving a backup specification
SQL Server-specific backup options
Specify SQL Server-specific backup options by clicking Advanced in the Application Specific
Options group box page.
Application-specific options
SQL Server backup options
Pre-exec
Specifies a command with arguments or a script started by sql_bar.exe
on SQL
Server before backup. Resides in the default Data Protector commands directory.
Only the filename must be provided in the backup specification.
Data Protector (10.00) Page 211 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Post-exec
Concurrent streams
Specifies a command with arguments or a script started by sql_bar.exe
on SQL
Server after backup. Resides in the default Data Protector commands directory.
Only the filename must be provided in the backup specification.
Sets the number of concurrent streams used to back up SQL Server databases from the replica to tape. Applicable for ZDB–to–tape and ZDB–to–disk+tape sessions.
Ignored for ZDB sessions.
Fast direct mode
Check database integrity
Performs data integrity validation before backup. If the check fails, the session completes with warnings.
SQL backup compression
Specify how Data Protector should handle the Microsoft SQL Server backup compression.
SQL Server settings (default)
Enable
Use SQL server settings (default)
Handles the backup compression according to the Microsoft SQL Server settings.
Executes the backup compression regardless of the Microsoft SQL Server settings.
Disable
Specifies that the backup compression should not be executed regardless of the
Microsoft SQL Server settings.
Exclude from backup
(available for standalone instance backup only)
Excludes specific databases from backup.
Availability Group Databases
Standalone databases
Excludes databases belonging to any availability group from backup.
Excludes all standalone databases from backup.
None (default) Does not exclude any database from backup.
Performs backup according to the
Microsoft SQL Server settings.
NOTE:
Do not use double quotes (" ") in object-specific pre-exec and post-exec commands.
Data Protector (10.00) Page 212 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Scheduling backups
You can run unattended ZDB at specific times or periodically. For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE Data Protector Administrator's Guide.
NOTE:
You cannot run ZDB to disk or ZDB to disk+tape if Track the replica for instant recovery is not selected in the backup specification.
Starting backup sessions
Interactive backups are run on demand. They are useful for urgent backups or restarting failed backups.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, and then MS SQL Server. Right-click the backup specification you want to use and select Start Backup.
3. Select Network load. For information on network load, click Help. Click OK
For ZDB sessions, the backup type is set to Full.
For ZDB to disk or ZDB to disk+tape, specify the Split mirror/snapshot backup option.
Starting interactive backups
Using the Data Protector CLI
To start ZDB to tape or ZDB to disk+tape, execute: omnib -mssql_list ListName
To start ZDB to disk, execute: omnib -mssql_list ListName -disk_only
Data Protector (10.00) Page 213 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration where
ListName
is the name of the backup specification. For more information on omnib
, see its man page.
Restore
Data Protector offers restore from backup media to the application system on LAN (standard restore), where you can select various restore options depending on your restore scenario, and instant recovery.
For more information, see the following sections.
Before you begin
l
Verify that the databases to be restored are not in use.
l
In an availability group configuration, restore to a different client and instance is mandatory. User must select the restore options with appropriate values for the fields 'Restore to another Client' and
‘Restore to another Instance’. Make sure that you do not select an availability group listener for the target client (as it is not supported), and that the selected SQL Server instance exists on the target client. Also make sure that the database which you selected for the restore does not belong to any availability group.
Standard restore
NOTE:
There is no need to create an empty database before restore, because the database and its files are generated automatically by SQL Server.
Proceed as follows using the Data Protector Manager:
1. In the Context List, click Restore.
2. In the Scoping Pane, expand Restore Objects, MS SQL Server, and then select the client
(backup system) from which you want to restore. A list of backed up objects is displayed in the
Results Area.
Data Protector (10.00) Page 214 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Selecting backup objects for restore
To select backup object-specific options, right-click the object and select Properties.
Selecting object-specific options
Data Protector (10.00) Page 215 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
In the Version tab, select the backup version (backup date) which you want to use for restore.
Select other restore options as appropriate. Note that some options are not available for restore of
data files. See Restore options, on the next page
for details.
Click OK.
4. In the Options property page, specify new locations for the databases, if you want to restore your data to a different client or instance.
IMPORTANT:
l
When you click Options, the cell is browsed for running SQL Server instances that can become target instances for restore. If no instances are found, Restore to another
instance is disabled and the message There are no instances on this client system is displayed.
l
Make sure that the specified SQL Server instance exists on the target client. Otherwise, restore fails.
Select one of the following Restore actions: l
Restore data . Select to restore the whole database. This option is selected by default.
l
Restore and display file list only . Select if you do not know the original filenames. In this case, the files backed up in a particular session are displayed.
l
Restore and display headers only . Select if you need specific details about backup. SQL
Server header information is displayed.
Data Protector (10.00) Page 216 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Restore options
5. In the Devices page, select the devices to be used for the restore.
For more information on how to select devices for a restore, see the HPE Data Protector Help index: “restore, selecting devices for”.
6. Click Restore MS SQL Server and then Next to select Report level and Network load.
NOTE:
Select Display statistical information to view the restore profile messages in the session output.
7. Click Finish to start the restore.
The statistics of the restore session, along with the message Session completed successfully
is displayed at the end of the session output.
Restore options
Microsoft SQL Server database restore options
Option Description
Backup version
Specifies the backup session from which the selected objects will be restored.
Data Protector (10.00) Page 217 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Option Description
Point-in-time restore
This option is only available for database objects.
Specifies a point in time to which the database state will be restored (you also need to select Backup version and set Stop at). After recovery, the database is in the state it was at the specified date and time.
Only transaction logs written before the specified date and time are applied to the database.
Stop at
This option is only available for database objects.
Specifies the exact time when the rollforward of transactions will be stopped. Therefore, to enable database recovery to a particular point in time, backup you restore from must be a transaction log backup.
You cannot use this option with NORECOVERY or STANDBY. If you specify Stop at time that is after the end of RESTORE LOG operation, the database is left in a non-recovered state (as if RESTORE LOG is run with NORECOVERY).
Restore only this backup
Full restore of the database
If you restored a database version and left it in a non-operational or standby state, you can subsequently restore differential or transaction log backups one by one, leaving each version non-operational to restore additional backups.
All necessary versions are restored, including the latest full backup, the differential backup (if one exists), and all transaction log backups from the differential up to the selected version.
Force restore over the existing database
Put database in single user mode - log off all users
Recovery completion state
Select this option if a database with the same name but a different internal structure already exists at the target Microsoft SQL Server instance.
If this option is not selected, the Microsoft SQL Server does not let you overwrite the existing database - the restore will fail.
If you are restoring a data file from the PRIMARY group to an existing database, you must specify the option at the data file level.
When using this option, make sure that the most recent logs are backed up before the restore.
Disconnects all users that are connected to the target Microsoft SQL
Server database and puts the database in the single user mode. Note that if the database is not in the simple recovery mode, the Force restore
over the existing database option should also be selected.
Enables selecting the database state after recovery. You may select from: l
Leaving the database operational. Once the last transaction log is
Data Protector (10.00) Page 218 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Option Description
Restore database with a new name
restored and the recovery completed, the database becomes operational.
l
Leaving the database non-operational after the last transaction log is restored. You may restore additional transaction logs one by one.
l
Leaving the database in read-only mode. You may restore additional transaction logs before the database is set to read-write mode.
This selection is only available for database objects.
This option is only available for database objects.
Restores the database under a different name. Specify the database logical filename and the destination filename (suboptions of Restore files
to new locations).
Restore files to new locations
Restores files to a new location. Specify the database logical filename and a destination target filename for the specified logical filename. Use this option to restore data to a different client, a different instance, or to make a database copy on the same client.
TIP:
To allow different restore scenarios, you can combine general restore options, such as Restore
database to another Microsoft SQL Server and Restore using a different device, with object-specific restore options, such as Point-in-time restore, Recovery completion state,
Force restore over the existing database.
Restoring to a different SQL Server instance or/and different SQL
Server
Prerequisites l
Both SQL Servers must have the same local settings (code page and sort order). This information is displayed in the session monitor for each backup.
l
The target SQL Server must be configured and reside in the same Data Protector cell as the original
SQL Server.
For the configuration procedure, see Creating ZDB specifications, on page 201
.
Procedure
1. Select the databases you want to restore and their versions.
2. Select the following: l
To restore to a different SQL Server client, select Restore to another client and the target client from the drop-down list.
l
To restore to a different SQL Server instance, select Restore to another instance. If there are no instances in the drop-down list, enter the instance name by yourself.
Data Protector (10.00) Page 219 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
IMPORTANT:
Make sure that the specified SQL Server instance exists on the target client.
Otherwise, restore fails.
3. Specify new database locations.
4. Start restore. See Restore, on page 214
.
In case of the following backup sequence: l
F(Italic) F T(Italic)(bold) T: To restore the version marked T, all the backup versions in italic are restored. Chain was constructed by LSN.
l
F F(Italic) T(Italic) T(Italic)(bold): To restore the version marked T, all the backup versions in italic are restored. Chain is constructed by time.
When Data Protector is not able to construct chain by LSNs, the old formula by time still applies as it was before. So the latest full backup, the latest differential backup and all transaction log backups after the last full or differential backup must be restored.
Instant recovery
See the HPE Data Protector Concepts Guide and HPE Data Protector Zero Downtime Backup
Administrator's Guide for general information on instant recovery.
Prerequisites l
If you restore user databases, put the databases offline:
1. Start SQL Server Enterprise Manager.
2. Selecting the database and click Action.
3. Select all tasks and take them offline.
l
If you restore system databases, put SQL Server offline by starting SQL Server Enterprise Manager, right-clicking SQL Server, and clicking Stop.
Perform instant recovery using the Data Protector Manager:
1. In the Context List, click Instant Recovery.
2. Expand MS SQL Server and select the backup session (replica) from which you want to restore.
By default, the database will be recovered until the last backed up transaction.
3. To recover user databases to a specific point in time:
a. In the Source property page, under Restore Objects, right-click a database and click
Properties.
b. In the Backup version drop-down list, select the required replica. The latest version is selected by default.
Select Point in time restore. From the Stop at drop-down list, select the point in time to which the transactions should be applied, and click OK. If no transaction logs are available, this option is disabled.
Data Protector (10.00) Page 220 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Point-in-Time restore
To restore the database under a different name, click Advanced and select Restore
database with new name. See Restoring database with a new name, below .
IMPORTANT:
If the logical filename and physical filename are not listed, add them to the list. Specify the same names as used for ZDB; otherwise, instant recovery fails.
Restoring database with a new name
Data Protector (10.00) Page 221 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
4. Click Restore MS SQL Server.
If you restore system databases, SQL Server displays errors because its services are offline.
Therefore, when restore completes, start SQL Server manually using SQL Server Enterprise
Manager.
Monitoring sessions
You can monitor currently running or view previous sessions in the Data Protector GUI. When you run an interactive session, the monitor window shows you the session progress. Closing the GUI does not affect the session.
You can also monitor sessions using the Monitor context from any Data Protector client with the User
Interface component installed.
For information on monitoring sessions, see the HPE Data Protector Help index: “viewing currently running sessions” and “viewing finished sessions”.
Troubleshooting
This section lists general checks and verifications, plus problems you might encounter when using the
Data Protector SQL Server integration. Start at Problems, on page 224
. If you cannot find a solution there, perform general checks and verifications.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
Data Protector (10.00) Page 222 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
For general ZDB, restore, and instant recovery related troubleshooting, see the HPE Data Protector
Zero Downtime Backup Administrator's Guide.
Before you begin
l
Make sure that the latest official Data Protector patches are installed. For details on how to verify this, see the HPE Data Protector Help index: “patches”.
l
See the HPE Data Protector Product Announcements, Software Notes, and References for general
Data Protector limitations, as well as recognized issues and workarounds.
l
See https://softwaresupport.hpe.com/manuals for an up-to-date list of supported versions, platforms, and other information.
Checks and verifications
If your configuration, backup, or restore failed: l
Check that SQL Server services are running.
l
Examine system errors reported in debug.log
on the SQL Server client.
Additionally, check errorlog and VDI.log
files in the MSSQL\log directory.
l
Make a test filesystem backup and restore of the problematic client. For information, see the HPE
Data Protector Help.
l
Check that every SQL Server used with Data Protector has the MS SQL Integration component installed.
l
Connect to SQL Server via SQL Server Enterprise Manager using the same login ID as you specified in the Data Protector Configuration dialog box.
l
Perform a database backup using SQL Server Enterprise Manager. If the backup fails, fix any SQL
Server problems, and then perform a backup using Data Protector.
Additionally, if your backup failed: l
Verify the configuration file to check if the Cell Manager is correctly set on SQL Server.
l
If you do not see the SQL Server instance as the application database when creating a backup specification, enter the instance name yourself. When “not-named instance” is not displayed, insert the
DEFAULT
string.
l
If Data Protector reports that the integration is properly configured, verify that the SQL Server user has appropriate rights to access the required databases.
During master database restore, the following error occurs when executing an SQL statement:
Error has occurred while executing an SQL statement.
Error message: 'SQLSTATE:[42000] CODE:(3108) MESSAGE:[Microsoft]
[ODBC SQL Server Driver][SQL Server]To restore the master database, the server must be running in single user mode. For information on starting in single user mode, see "How to: Start an Instance of SQL
Server (sqlservr.exe)" in Books Online.
Note that this behavior is expected when the master database is not restored in single user mode.
Data Protector (10.00) Page 223 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Problems
Problem
The integration is properly configured but the database backup fails after a timeout
l
With an error similar to:
[Warning] From: [email protected] "SQLSRV"
Time: 7/29/2011 8:19:22 PM
Error has occurred while executing SQL statement.
[Microsoft][ODBC SQL Server Driver][SQL Server]Backup or restore operation terminating abnormally.'
[Critical] From: [email protected] "SQLSRV"
Time: 7/29/11 8:19:24 PM
Received ABORT request from SM => aborting l
SQL Server error log contains an entry similar to:
2011-07-29 20:19:21.62 kernel
BackupVirtualDeviceSet::Initialize: Open failure on backup device 'Data_Protector_master'.
Operating system error -2147024891(Access is denied.).
l
SQL Server VDI.LOG
file contains an entry similar to:
2011/07/30 13:19:31 pid(2112)
Error at BuildSecurityAttributes: SetSecurityDescriptorDacl
Status Code: 1338, x53A Explanation: The security descriptor structure is invalid.
SQL Server service and Data Protector Inet are running under different accounts. The integration cannot access SQL Server due to security problems.
Action
Restart the Data Protector Inet service under the same account as the SQL Server service is running.
Problem
Backup fails with "The object was not open"
When backing up Microsoft SQL Server databases, the session fails with an error similar to the following:
[Critical]From : OB2BAR_Main@wemaoldb2dr "Aolins" Time:11/12/2011 02:01:34 AM
Microsoft SQL Server reported the following error during login : The object was not open
The error may appear if the SQL Server Browser service is not running.
Action
Proceed as follows:
Data Protector (10.00) Page 224 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
1. Start the SQL Server Browser service.
2. Start a new backup session.
Problem
Backup fails if the appropriate drive letter on the backup system does not exist
Backup fails with an error, similar to:
[Major] From: [email protected] "" Time: 02-Feb-11 14:07:54
Filesystem \\.\Volume{ef58fe0e-b2b8-11db-aa08-000802804af6} could not be mounted to Q:\Program Files\Microsoft SQL Server\MSSQL.1\MSSQL\DATA\.
([87] The parameter is incorrect. ).
Backup fails because SSE agent tries to mount the filesystem to the drive letter which does not exist on the backup system. The drive letter must be the same as on the application system. SSE or SMI-S agents always mount the filesystems to the same drive letters on the backup systems as if the ZDB_
PRESERVE_MOUNTPOINTSomnirc option is set to 1 .
Action
For source volumes to be successfully replicated, create the same drive letters on the backup system as they are used on the application system for mounting the source volumes.
Problem
Database is left in unrecovered state after “Invalid value specified for STOPAT parameter” is reported
The database remains in an unrecovered state as if the
RESTORE LOG
operation was run with Leave the
database non-operational.
Action
Recover the database to the latest point in time using SQL Query Analyzer:
RESTORE DATABASE database_name WITH RECOVERY
After the recovery, additional transaction logs cannot be applied.
Problem
Transaction logs cannot be restored from tape
The recovery completed successfully and the database was put in norecovery state, but transaction logs cannot be restored from tape.
Action
Recover the database to the state of ZDB to disk using the SQL Query Analyzer:
RESTORE DATABASE database name WITH RECOVERY
After the recovery, additional transaction logs cannot be applied.
Problem
Instant recovery of SQL Server databases fails
Data Protector (10.00) Page 225 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
If the SQL Server service is offline prior to the instant recovery, instant recovery of the SQL Server databases fails.
The following errors are displayed:
[Critical] From: [email protected] "(DEFAULT)" Time: 4/9/2011 7:01:42 PM
Microsoft SQL Server reported the following error during login:
The object was not open.
[Warning] From: [email protected] "(DEFAULT)" Time: 4/9/2011 7:01:42 PM
[152:9208] Data Protector is probably not configured for use with SQL Server on this host.
Action
Perform one of the following: l
Set the Data Protector omnirc option OB2_SQLRESTORE_STARTSRV , which starts the SQL Server service prior to the recovery of SQL databases, to 1 .
During the master database restore, the following error is displayed:
RESTORE master with SNAPSHOT is not supported.
Note that this behavior is expected. No further steps are needed after the instant recovery.
l
Restart SQL Server instance services after instant recovery completes. If restarting services does not automatically start the recovery of all system databases, start the SQL Server instance in single user mode and manually start the recovery of the master database. Follow the same procedure for other system databases. At the end, restart SQL Server instance services.
Problem
Restore to another client in the Data Protector cell not configured for use with SQL Server fails
Action
Configure the SQL integration on this client (see Configuring the integration, on page 195
).
Problem
Database is left in unrecovered state after restore completed successfully
If you set the time for
Stop at
beyond the end of the
RESTORE LOG
operation, the database remains in the unrecovered state as if the
RESTORE LOG
operation was run with
Leave the database nonoperational .
Action
Recover the database to the latest point in time by using the SQL Query Analyzer:
RESTORE DATABASE database_name WITH RECOVERY
After the recovery, additional transaction logs cannot be applied.
Problem
Instant recovery of a Microsoft SQL Server database configured on a Microsoft Cluster Server system fails with "The physical filename may be incorrect"
Data Protector (10.00) Page 226 of 413
Zero Downtime Backup Integration Guide
Chapter 3: Data Protector Microsoft SQL Server ZDB integration
Instant recovery of the Microsoft SQL Server data in an HPE Business Copy P9000 XP configuration on a Microsoft Cluster Server system fails with the following error:
[Microsoft][ODBC SQL Server Driver][SQL Server]Device activation error.
The physical file name '<Data/Log filename>' may be incorrect.
Action
Perform the following steps:
1. Using Microsoft SQL Server Enterprise Manager, detach the Microsoft SQL Server database that you want to recover.
2. Using Cluster Administrator, take the Microsoft SQL Server Disk resource offline.
3. On the application system, configure the ZDB_TAKE_CLUSRES_ONLINEomnirc option.
For details, see ???.
4. Start instant recovery.
5. When the message Please, take MS SQL cluster resources online appears in the Data
Protector GUI, bring the Microsoft SQL Server cluster resources online using Cluster
Administrator.
Problem
Database restore fails
The restore session aborts with a major error similar to:
Error has occurred while executing a SQL statement.Error message: 'SQLSTATE:[42000]
CODE:(3159) MESSAGE:[Microsoft][ODBC SQL Server Driver][SQL Server]The tail of the log for the database "test2" has not been backed up. Use BACKUP LOG WITH NORECOVERY to backup the log if it contains work you do not want to lose. Use the WITH REPLACE
or WITH STOPAT clause of the RESTORE statement to just overwrite the contents of the log. SQLSTATE:[42000] CODE:(3013) MESSAGE:[Microsoft][ODBC SQL Server Driver]
[SQL Server]RESTORE DATABASE is terminating abnormally.
Action
To solve the problem perform a transaction log backup to obtain the most recent transaction logs.
Data Protector (10.00) Page 227 of 413
Chapter 4: Data Protector Microsoft
Exchange Server 2010+ ZDB integration
Introduction
This chapter explains how to configure and use the Data Protector Microsoft Exchange Server 2010 ZDB integration, where Data Protector integrates with Microsoft Exchange Server 2010 or Microsoft Exchange
Server 2013 (hereinafter, both Exchange Servers are called Microsoft Exchange Server, unless differences are pointed out). It describes concepts and methods you need to understand to back up and restore Microsoft
Exchange Server 2010 mailbox and public folder databases or Microsoft Exchange Server 2013 mailbox databases (databases).
Both standalone environments and Database Availability Group (DAG) environments are supported.
The Data Protector Microsoft Exchange Server 2010 integration is based on the Volume Shadow Copy
Service (VSS) technology. For details on VSS concepts, see the HPE Data Protector Integration Guide for
Microsoft Volume Shadow Copy Service.
Backup
During backup, databases can be used actively (online backup). In DAG environments, you can back up active and/or passive database copies.
As Microsoft Exchange Server, ZDB disk array, and VSS are involved, you can specify different kinds of backup types: l
Microsoft Exchange Server backup types l
VSS backup types l
ZDB backup types
You can select among the following Microsoft Exchange Server backup types: l
Full l
Copy l
Incremental l
Differential
You can select among the following ZDB backup types: l
ZDB-to-disk l
ZDB-to-disk+tape l
ZDB-to-tape
You can select among the following VSS backup types: l
Local or network backup l
VSS transportable
For details on the backup types, see Backup types, on page 235 .
Data Protector (10.00) Page 228 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore
You can restore Microsoft Exchange Server databases using standard restore or instant recovery.
During restore, each database can be restored using a different restore method. The following methods are available: l
Repair all passive copies with failed status l
Restore to the latest state l
Restore to a point in time l
Restore to a new mailbox database l
Restore files to a temporary location
This chapter provides information specific to the Microsoft Exchange Server 2010 integration. For additional limitations, see the HPE Data Protector Product Announcements, Software Notes, and
References. For general Data Protector procedures and options, see the HPE Data Protector Help.
Integration concepts
Data Protector integrates with Microsoft Exchange Server through the Data Protector Microsoft
Exchange Server integration agent, which channels communication between the Data Protector
Session Manager and the clients in the Microsoft Exchange Server environment. The agent communicates with Microsoft Exchange Server through the Microsoft Exchange Management Shell and uses VSS to back up data.
Supported environments
Data Protector supports Microsoft Exchange Server Database Availability Group environments (DAG
environments) as well as environments with standalone Microsoft Exchange Server systems
(standalone environments).
Standalone environments
In a standalone Microsoft Exchange Server environment, each Microsoft Exchange Server system stands on its own.
In one session, you can back up databases from only one Microsoft Exchange Server system. Data
Protector sends backup and restore requests directly to the Microsoft Exchange Server system.
Data Protector (10.00) Page 229 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Standalone environment (example)
DAG environments
In a DAG environment, Data Protector communicates with the DAG using one of the Microsoft
Exchange Server systems (the one that is currently active in the environment). All backup and restore requests are sent there.
In one session, you can back up active and/or passive database copies from different Microsoft
Exchange Server systems that belong to the same DAG.
DAG environment (example)
Data Protector (10.00) Page 230 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
In DAG environment (example), on the previous page
, active databases are shaded in blue.
If a database has multiple passive copies, you can specify which particular passive copy you want to back up, using one of the following backup policies: l minimize the number of hosts l lowest activation preference l highest activation preference l shortest replay lag time l longest replay lag time l longest truncation lag time
You can also specify from which Microsoft Exchange Server systems database copies should not be backed up.
Data Protector (10.00) Page 231 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For a brief description of the activation preference number, replay lag time, and the truncation lag time,
see Microsoft Exchange Server parameters in DAG environments, below
.
Microsoft Exchange Server parameters in DAG environments
Parameter
Activation preference number
Description
The activation preference number determines which passive copy is activated if multiple passive copies meet the same criteria; the copy assigned the lowest activation preference number is activated.
Replay lag time
Truncation lag time
The ReplayLagTime parameter plays a role when synchronizing a passive copy with the active copy. As soon as a log file at the active copy side is filled up, it is copied to the passive copy side. By default, the newly copied log is also applied to the passive copy database files.
However, if the passive copy ReplayLagTime parameter is set to a value greater than 0 , the log is applied with a lag, creating a lagged database copy.
The maximum value is 14 days.
The
TruncationLagTime
parameter specifies how long the Microsoft
Exchange Replication service waits before truncating log files that have already been applied to the database files.
The maximum value is 14 days.
Configuring the integration
Prerequisites
l
Ensure that you have correctly installed and configured the Microsoft Exchange Server environment.
o
For supported versions, platforms, devices, and other information, see the latest support matrices at https://softwaresupport.hpe.com/manuals .
o
On Microsoft Exchange Server, install .NET Framework 3.5.1.
NOTE:
On Windows Server 2012, the installation of .NET Framework 3.5.1 is done manually and not by default. o
For information on installing, configuring, and using Microsoft Exchange Server, see the
Microsoft Exchange Server documentation.
o
In a Microsoft Exchange Server 2010 environment, if you intend to use the restore method
Restore to a point in time, make sure that you have Microsoft Exchange Server 2010 SP1 installed.
Data Protector (10.00) Page 232 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration l
If you intend to run Incremental and Differential backup sessions, make sure that circular logging is disabled.
If you plan to run instant recovery sessions, it is advisable to keep Microsoft Exchange Server databases on separate storage volumes. Also, keep a database's files ( .edb
and .log
) on separate
storage volumes. See Where to keep database files, below . Such a configuration provides better
restore granularity.
Ensure that storage volumes on different Microsoft Exchange Server systems are the same size.
Otherwise, you may experience problems during copy-backup instant recovery sessions.
Where to keep database files
l
Ensure that you have correctly installed Data Protector. On how to install Data Protector in various architectures, see the HPE Data Protector Installation Guide.
Ensure that the following Data Protector components are installed on all Microsoft Exchange Server systems: o
MS Exchange Server 2010+ Integration o
MS Volume Shadow Copy Integration o
The appropriate Data Protector disk array agent
NOTE:
For VSS transportable backup sessions, the
MS Volume Shadow Copy Integration component and the appropriate Data Protector disk array agent must also be installed on the backup systems.
In DAG environments, the DAG virtual system (host) must also be imported to the Data Protector
Cell. On how to import a client to a Data Protector Cell, see the HPE Data Protector Help index:
"importing, client systems”.
l
For limitations, see “Limitations and recommendations” in the HPE Data Protector Product
Announcements, Software Notes, and References.
Data Protector (10.00) Page 233 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Limitations
l
Due to incompatibility between Microsoft Exchange Server versions, backup objects belonging to a particular Exchange Server version cannot be restored to Data Protector clients on which a different
Exchange Server version is installed.
Before you begin
l
Configure devices and media for use with Data Protector.
l
To test whether a Microsoft Exchange Server system and the Cell Manager communicate properly, configure and run a Data Protector filesystem backup and restore on every Microsoft Exchange
Server client in your environment.
Configuring user accounts
Windows domain user account for backup and restore sessions
Backup and restore sessions are started by the Data Protector Inet service, which by default runs under the Windows local user account
SYSTEM
. Consequently, a backup or restore session is performed using the same user account.
However, you can specify that the Data Protector Inet service should use a different Windows domain user account to start a session: l
To perform a backup session under a different user account, specify the Specify OS user option
(see Specifying view type, on page 241
) when creating a backup specification.
l
To perform a restore session under a different user account, specify the User name and
Before you specify a different Windows domain user account, configure the user account as follows:
1. Grant the user appropriate permissions to back up and restore Microsoft Exchange Server databases.
2. Add the user to the Data Protector admin or operator user group. For details on adding users, see the HPE Data Protector Help index: ”adding users”.
3. Save the user and its password to a Windows Registry on the Microsoft Exchange Server system on which you plan to start the integration agent ( e2010_bar.exe
). To save the user account, use the Data Protector omniinetpasswd or omnicc command.
NOTE:
The user account saved in the Windows Registry will be used by the Data Protector
Inet
service when needed.
For details on setting accounts for the Inet service user impersonation, see the HPE Data
Protector Help index: “Inet user impersonation”.
Data Protector (10.00) Page 234 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Example
To save the user jane
from the domain
HPE
and with the password mysecret
to a Windows
Registry, log on to the Microsoft Exchange Server system and execute the following command: omniinetpasswd –add jane@HPE mysecret
User account for executing Exchange Management cmdlet operations
In the Microsoft Exchange Server 2013 environment, you need user credentials with specific Exchange
Management Roles assigned to create a remote runspace for executing the Exchange Management cmdlet operations remotely. These operations are executed as part of Microsoft Exchange Server backup and restore operations.
Configure your user account with the following Exchange privileges: l
As a member of the Organization Management role group.
l
As a member of the Discovery Management role group.
l
As a member of the Administrators group of the Microsoft Exchange Server system on which the integration is installed.
Configure a valid Exchange domain user account, when creating a backup specification. The user account is saved in the user credentials specific configuration file located in the Data_Protector_ program_data\Config\Server\Integ\Config\E2010 directory and named by the domain name. The saved user credentials will be used by Data Protector when needed.
For details, see Specifying view type, on page 241
.
For information on the Exchange Management cmdlet operations, see the Microsoft Exchange Server documentation.
Backup
When you back up a Microsoft Exchange Server database, the following files are automatically backed up: l database files ( .edb
) l transaction logs (
.log
) l checkpoint files ( .chk
)
However, depending on the Microsoft Exchange Server backup type you select, not all files are always
backed up. For details, see Microsoft Exchange Server backup types, on the next page
.
Backup types
As Microsoft Exchange Server, ZDB disk array, and VSS are involved, you can specify different kinds of backup types:
Data Protector (10.00) Page 235 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration l
Microsoft Exchange Server backup types l
ZDB backup types l
VSS backup types
Microsoft Exchange Server backup types
You can select among the following Microsoft Exchange Server backup types:
Backup types
Full
Backs up the database files (
.edb
), transaction logs (
.log
), and checkpoint files (
.chk
), and then truncates the transaction logs.
DAG environments: If multiple copies of a database are selected for backup, Data Protector first performs a Full backup of the passive copy that has the fewest logs applied to the database file, and then performs a Copy backup of all the remaining copies, with the active copy being backed up last. The copies are backed up sequentially due to a
Microsoft Exchange Server VSS writers limitation.
Copy
Incremental
Differential
Backs up the database files ( .edb
), transaction logs ( .log
), and checkpoint files ( .chk
), without truncating the transaction logs.
Backs up the transaction logs ( .log
) that have been created since the last Full or Incremental backup, and then truncates the transaction logs.
DAG environments: If multiple copies of a database are selected for backup, Data Protector backs up the transaction logs of only one copy
(one of the passive copies is selected).
Backs up the transaction logs ( .log
) that have been created since the last Full backup, without truncating the transaction logs.
NOTE:
An incremental or differential backup of a database cannot be performed: l
If a full backup has not been performed.
l
If an incremental backup is started just after a differential backup has been performed, or the other way around.
l
If Microsoft Exchange Server circular logging is enabled.
ZDB backup types
You can select among the following ZDB backup types: l
ZDB-to-disk l
ZDB-to-disk+tape l
ZDB-to-tape
Data Protector (10.00) Page 236 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
NOTE:
For Microsoft Exchange Server Incremental and Differential backup types, only the ZDB-totape backup type is available.
VSS backup types
You can select among the following VSS backup types: l
Local or network backup l
VSS transportable
For details, see the HPE Data Protector Integration Guide for Microsoft Volume Shadow Copy Service.
Backup parallelism
l
During a backup session, copies of different databases are backed up in parallel, however, copies of the same database are not, due to a Microsoft Exchange Server VSS writers limitation.
l
If multiple backup sessions that intend to back up the same database are started in parallel, only the session that first locks the database can back up the database; the other sessions cannot. In DAG environments, this also applies if backup sessions intend to back up different copies of the same database; only the session that first locks the database (that is, all its copies) can back up the database copies; the other sessions cannot.
NOTE:
This behavior ensures that the construction of a restore chain is valid. For example, suppose that several full backup sessions that intend to back up the same database are started in parallel. If all the sessions backed up the database, it might happen that the session given the latest Session ID is not the one that backed up the database last. For details on restore
chains, see Restore chain, on page 255
.
Replica rotation in DAG environments
Data Protector Microsoft Exchange Server 2010 integration enables you to perform ZDB sessions in multi-system environments (DAG environments). This brings some changes to the existing replica rotation functionality.
In standalone environments, the replica rotation functionality works as it used to; it limits the number of backups that are kept in the Data Protector IDB database for instant recovery purposes. For example, if the Number of replicas rotated option is set to 1 , only one backup session is available for instant recovery at a time. If you start another backup session (with the same backup specification), backup storage volumes created in the previous backup session are deleted before new ones are created and the previous session is removed from the Instant Recovery context.
In DAG environments, a database can be backed up from different systems in different sessions. This introduces the changes. For example, you want to back up the database DB1 , which is active on node1.company.com
, and the database DB2 , which is active on node2.company.com
. Suppose the
Number of replicas rotated option is set to 1 and your backup policy is such that the active copy is always backed up. After the backup, the Data Protector VSSDB database contains the following entries:
Data Protector (10.00) Page 237 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Backup 1 (
2013/10/05–1
): l
2013/10/05-1:node1.company.com
(containing DB1 ) l
2013/10/05-1:node2.company.com
(containing
DB2
)
Suppose that a failover occurs and the database DB1 becomes active on node2.company.com
. You start another backup session. Now both databases are backed up from node2.company.com
. As a result, the VSSDB database contains the following entries:
Backup 2 (
2013/10/05–2
): l
2013/10/05-1:node1.company.com
(containing DB1 ) l
2013/10/05-2:node2.company.com
(containing
DB1
and
DB2
)
The entry 2013/10/05-1:node2.company.com
is no longer in the VSSDB database as it has been rotated out (the corresponding backup storage volumes have been deleted) due to the replica rotation functionality.
NOTE:
Entries are rotated per system and not per session. As a result, entries that are created in the same session can be rotated out at different points in time (that is, in different sessions).
Now another failover occurs and both databases become active on node1.company.com
. You start another session and both databases are backed up from node1.company.com
. As a result, the VSSDB database contains the following entries:
Backup 3 (
2013/10/05–3
): l
2013/10/05-3:node1.company.com
(containing DB1 and DB2 ) l
2013/10/05-2:node2.company.com
(containing
DB1
and
DB2
)
Note that the entry 2013/10/05-1:node1.company.com
has been rotated out as well. Since both parts created in the session 2013/10/05-1 have been rotated out, the session 2013/10/05-1 can no longer be used for instant recovery (it is removed from the Instant Recovery context).
If you go to the Instant Recovery context and select the session created in Backup 1 after you have performed Backup 2, the source page incorrectly shows that both databases can be restored. The entry
2013/10/05-1:node2.company.com
(containing DB2 ) has been rotated out in Backup 2. If you select
DB2 for restore and start instant recovery, the session fails. Therefore, make sure to restore databases from sessions for which the necessary entries in the VSSDB database still exist.
Backup considerations
l
Backup strategy:
Choose one of the following strategies to back up your data: o
Full o
Full, Incremental, Incremental, ...
o
Full, Differential, Differential, ...
o
Full, Copy, Incremental, ..., Copy, Incremental, ...
Data Protector (10.00) Page 238 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
IMPORTANT:
An Incremental backup session cannot be followed by a Differential backup session, nor the other way around. You must first run a Full backup session.
l
Active copies as opposed to passive copies:
There is no difference between the active and passive copy, except in the currently active log file (at the active copy side), which is not copied to the passive copy side until the file is filled up (that is, reaches 1 MB). Consequently, if you back up a passive copy, the transactions in the currently active log file are not included.
l
Lagged database copies:
Backing up a lagged database copy is equivalent to backing up a non-lagged database copy. If you restore from the backup of a lagged database copy, files are not only restored, but logs are also applied to the database file, returning the database to its most recent state. However, restoring the logs and applying them to the database file is time-consuming and, therefore prolongs the restore session. Also note that you need enough disk space to restore all the necessary logs.
On the other hand, restoring from the backup of a lagged database copy enables you to restore the database to a point in time before the backup was taken. Restore the database without performing database recovery and mounting. Then remove unwanted logs, and finally recover and mount the database.
l
Public folders :
In the Microsoft Exchange Server 2010 environment, backup of Microsoft Exchange Server public folders with activated replication is not supported.
l
Concurrent backup sessions:
Backup sessions that back up the same database cannot run in parallel.
Object operations considerations
l
Object copy and object verification
When copying or verifying Microsoft Exchange Server objects you need to select all Data Protector backup objects created in the same session. To make sure that you do not select only a few objects from the session, the Data Protector GUI does not list Microsoft Exchange Server backup objects for interactive object copy or object verification sessions in the Objects scope of the Object
Operations context.
Use the Session or the Media scope instead.
Creating backup specifications
Create a backup specification using the Data Protector GUI (Data Protector Manager).
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click MS Exchange 2010+ Server, and click Add Backup.
3. In the Create New Backup dialog box, specify Backup type (VSS backup type). For details, press
F1. Click OK.
Data Protector (10.00) Page 239 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
4. In Application system, select the Microsoft Exchange Server system that you want to back up.
In a DAG environment, select the DAG virtual system or a Microsoft Exchange Server system.
NOTE:
The Application system drop-down list contains all clients that have the Data Protector
MS Exchange Server component installed. In a DAG environment, the list contains also the DAG virtual system (host).
The backup session (that is, the integration agent e2010_bar.exe
) will be started on the client that you specify here. If you select a DAG virtual system, the integration agent is started on the currently active Microsoft Exchange Server node. To find out which node is
currently active, see Restore, on page 251 .
NOTE:
In the Microsoft Exchange Server 2010 environment, to back up public folders residing on a
Microsoft Exchange Server system that is a part of a DAG environment, select the
Microsoft Exchange Server system and not the DAG virtual system (host). If you select the DAG virtual system, you can back up only databases that belong to the DAG. The
Microsoft Exchange Server public folders database is not the part of it.
Depending on the VSS backup type you selected, specify the following: l
If you selected Local or network backup, in Provider, select Use hardware provider.
l
If you selected VSS transportable backup, specify Backup system.
Specify ZDB-specific options.
For details, press F1 or see the HPE Data Protector Integration Guide for Microsoft Volume
Shadow Copy Service.
NOTE:
In an HPE P9000 XP Disk Array Family resync mode environment, the maximum number of replica storage volumes (S-VOL) that can be created for a given storage volume (P-VOL) is limited by the hardware provider configuration — MU range (maximum is 3). To enable performing of incremental and differential sessions, the Number of replicas rotated option in the backup specification must be set to one less than the MU range. Thus, one replica storage volume is kept free for incremental and differential backup sessions.
Click Next.
5. In the Configure MS Exchange 2010+ Server dialog box, provide the domain, username, and password to browse or backup or recover the Exchange server.
Data Protector (10.00) Page 240 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
User credentials for executing Exchange Management cmdlet operations
Click OK.
6. The MS Exchange 2010+ Server is configured. Exit the GUI or proceed with creating the backup
7. If you selected the DAG virtual system (host), specify View Type to define how Microsoft
Exchange Server databases should be organized in the next page (source page):
By
Role
All databases in the DAG are displayed.
By
Client
All clients in the DAG are displayed, together with all the databases (active or passive) residing on them. Active databases have the label (active) appended at the end.
Passive databases have no label.
Specifying view type
Data Protector (10.00) Page 241 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For information on the User and group/domain options, press F1.
NOTE:
If no valid user credentials for executing the Exchange Management cmdlet operations remotely are specified, the Microsoft Exchange Server configuration dialog box is displayed.
8. Select which Microsoft Exchange Server databases to back up.
NOTE:
DAG environments:
In a single session, you can back up one of the following: l multiple databases, but only one copy of each l a single database, but multiple copies of it
Selecting databases (DAG environment – by role)
Data Protector (10.00) Page 242 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Selecting databases (DAG environment – by client)
Selecting databases (standalone environment)
Data Protector (10.00) Page 243 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
9. The following applies in DAG environments if you selected the By Role view type.
Specify the backup policy options.
Backup policy options
Data Protector (10.00) Page 244 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For details, see Backup policy options, on the next page .
10. Select the devices to use for the backup.
To specify device options, right-click the device and click Properties. Specify the number of parallel backup streams in the Concurrency tab and the media pool to use.
Click Next.
11. Set backup options.
Application-specific option
Click Next.
12. Click Save As to save the backup specification, specify a name and a backup specification group. Optionally, you can click Save and Schedule to save, and then schedule the backup specification. For more information on how to create and edit schedules, see Scheduler in Data
Protector in HPE Data Protector Administrator's Guide.
Data Protector (10.00) Page 245 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
TIP:
Backup policy options
Options Description
Back up active database
Back up passive copy
Back up all copies
If this option is selected, the active copy is backed up.
If this option is selected, a passive copy is backed up. If a database has multiple passive copies, specify which particular copy you want to back up, using one of the following policies:
minimize the number
of hosts (default)
If this option is selected, the minimum number of clients is involved in the backup. For example, if databases to be backed up have each a passive copy residing on the same client, they are all backed up from this client (and not one database from one client and another database from another client).
lowest/highest activation preference
If this option is selected, the database copy with the lowest/highest activation preference number is backed up.
shortest/longest replay lag time
If this option is selected, the database copy with the shortest/longest replay lag time is backed up.
longest truncation lag time
If this option is selected, the database copy with the longest truncation lag time is backed up.
For a brief description of the activation preference number, replay lag time and
transaction lag time parameters, see Microsoft Exchange Server parameters in
DAG environments, on page 232 . For details, see the Microsoft Exchange Server
documentation.
Use active copy for backup if no passive copies are available
Available if Back up passive copy is selected. If this option is selected, the active copy is backed up when no passive copy is available.
Available if only one database is selected for backup.
If this option is selected, all copies (active and passive) are backed up. This is useful when you create
ZDB-to-disk
or
ZDB-to-disk+tape
backups (that is, backups that can be used for instant recovery). If multiple copies are backed up, during instant recovery, multiple copies can be restored, as each copy has its own
Data Protector (10.00) Page 246 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Options Description
replica storage volumes to be restored from. For details, see Instant recovery in
DAG environments, on page 255 .
When you create a
ZDB-to–tape
backup, it is enough that a single copy is backed up; you can restore different copies of a database from the ZDB-to-tape backup of a single copy.
Creates a list of clients. The database copies that reside on these clients are not backed up.
Exclude clients from backup
Application-specific backup options
Options Description
Pre-exec , Post-exec
Perform consistency check
[-exch_check
[
-exch_throttle Value
] |
-exch_checklogs]
Specifies which command line to run on a Microsoft
Exchange Server system before ( pre-exec ) or after
( post-exec ) the backup.
The command line is executed only on the Microsoft
Exchange Server system on which the backup session is started (that is the system on which the Data Protector
Microsoft Exchange Server integration agent e2010_ bar.exe
is started).
Type only the name of the command and ensure that the command is located in the default Data Protector commands directory on the same system. Do not use double quotes.
DAG environments: If you selected the DAG virtual system (host) in the Application system option, ensure that the command is located on the currently active node.
To find out which Microsoft Exchange Server node is
currently active, see Restore, on page 251
.
If this option is selected, Microsoft Exchange Server checks the consistency of a database's backup data. If this option is not selected, the session finishes earlier, but the backup data consistency is not guaranteed.
The check is performed on the replica storage volumes after the backup data is created. If the data is found corrupt, the replica storage volumesare discarded and the database backup fails.
Default: selected
Data Protector (10.00) Page 247 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Options Description
If the Check log files only option is selected, only the backup data of the log files is checked, which is enough for Microsoft Exchange Server to guarantee data consistency.
Default: selected
By default, the consistency check is I/O intensive, which can negatively affect disk performance. The Throttle
check for 1 second option throttles down the consistency check of the database file .edb
to lessen impact on the disk performance. Specify after how many input/output operations the check should stop for one second.
This option is not available if only the log files are checked.
Default: not selected
Modifying backup specifications
To modify your backup specification, click its name in the Scoping Pane of the Backup context.
In the Microsoft Exchange Server 2013 environment, in the Source page, you can change the
Exchange domain user credentials for executing the Exchange Management cmdlet operations remotely by right-clicking the selected backup object and clicking Configure. You can also validate your configuration by clicking Check configuration.
Click other desired tabs, and apply the changes.
NOTE:
To see all databases in the source page, select All in the Show option. In a DAG environment, this not only shows all databases, but also updates the current status of databases (active or passive).
To see only the databases you selected for backup or excluded from backup, select Selected in the Show option. If any databases are not displayed, it means that they are not excluded from the backup specification and will be backed up. When a new database is added on the client selected for backup, it will be automatically included to the backup specification.
Scheduling backup sessions
You can schedule a backup session to start automatically at specific times or periodically. For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE Data
Protector Administrator's Guide.
Data Protector (10.00) Page 248 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Previewing backup sessions
Preview the backup session to test it. You can use the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications and then MS Exchange 2010+ Server.
Right-click the backup specification you want to preview and click Preview Backup.
3. Specify Backup type and Network load. Click OK.
The message Session completed successfully is displayed at the end of a successful preview.
Using the Data Protector CLI
1. Log in to the Cell Manager or to any client with the Data Protector User Interface component
.
2. Execute the following command: omnib -e2010_list BackupSpecificationName -test_bar
What happens during the preview?
The following are tested: l
Communication between the Microsoft Exchange Server system on which the backup session is started and the Cell Manager l
If each selected database has at least one copy available for backup after the Backup policy options and Client filtering options have been applied (this applies to backup specifications that contain backup policy options) l
If the selected databases are ready to be backed up (that is, they should not be dismounted, suspended, or in a failed state)
Starting backup sessions
Interactive backups are run on demand. They are useful for urgent backups or restarting failed backups.
To start a backup, use the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, and then MS Exchange 2010+ Server.
Right-click the backup specification you want to use and click Start Backup.
3. Specify Backup type and Network load. Click OK.
Data Protector (10.00) Page 249 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
The message
Session completed successfully
is displayed at the end of a successful backup session.
Using the Data Protector CLI
1. Log in to the Cell Manager or to any client that has the Data Protector User Interface
.
2. Execute the following command: omnib -e2010_list BackupSpecificationName [-barmode E2010Mode] [LIST_OPTIONS
] where
E2010Mode
is one of the following: full|copy|incr|diff
The default is full .
For
ListOptions
, see the omnib
man page or the HPE Data Protector Command Line Interface
Reference.
Examples
To start a Full backup using the backup specification MyDatabases , execute: omnib -e2010_list MyDatabases -barmode full
To start a Differential backup using the same backup specification, execute: omnib -e2010_list MyDatabases -barmode diff
Backup objects
For each database (copy), Data Protector creates the following backup objects: l
Database file object
o
:/Microsoft Exchange Writer (Exchange Information Store)/Microsoft
Information Store/DBID/File [MSVSSW-APP]
(standalone database or active copy) o
:/Microsoft Exchange Writer (Exchange Replication Service)
/Microsoft Information Store/DBID/File [MSVSSW-APP]
(passive copy) l
Log file object
o
:/Microsoft Exchange Writer (Exchange Information Store)/Microsoft
Information Store/DBID/Logs [MSVSSW-APP]
(standalone database or active copy) o
:/Microsoft Exchange Writer (Exchange Replication Service)
/Microsoft Information Store/DBID/Logs [MSVSSW-APP]
(passive copy) l
Database object
Data Protector (10.00) Page 250 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
The database object contains information needed to construct the restore chain. For details on
restore chains, see Restore chain, on page 255 .
l
VSS metadata object
/BackupSession/Metadata [MSVSSW-APP]
Information on whether the objects were successfully backed up or not is saved in the Data Protector
.
Restore
You can restore Microsoft Exchange Server data by performing a standard restore or instant recovery session.
For details, see Standard restore, on page 257
and Instant recovery, on page 267 .
IMPORTANT:
After you have restored a database, start a Full backup session for the database. Otherwise, the subsequent Incremental and Differential backup sessions will fail.
Considerations
l
A Microsoft Exchange Server database that was backed up using the Data Protector Microsoft
Volume Shadow Copy Service integration cannot be restored using the Data Protector Microsoft
Exchange Server 2010 integration, nor the other way round.
Restore methods
There are various reasons for restoring a Microsoft Exchange Server database. Here are some examples: l
The database has become corrupt.
l
The synchronization between an active and passive database copy is broken, but you want to avoid reseeding the passive copy, or simply the resume operation does not work.
l
The database needs to be restored to a different point in time.
l
The database's backup data needs to be restored for investigation purposes.
l
The database's backup data needs to be restored to a recovery database in order to extract individual mailboxes or mailbox files.
l
The database's backup data needs to be restored to a dial tone database.
To suit your needs, Data Protector offers different restore methods. You can choose among the following: l
Repair all passive copies with failed status
l
Restore to the latest state
l
Restore to a point in time
Data Protector (10.00) Page 251 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration l
Restore to a new mailbox database
l
Restore to a temporary location
You can specify different restore methods for different databases in the same session.
NOTE:
The first three methods restore backup data to the original database and are therefore only available if the original database still exists. The last two methods restore backup data to a new location.
Repair all passive copies with failed status
This method is available only for databases that are part of a DAG. It is useful if some of a database's passive copies become corrupt, acquiring the status Failed or FailedAndSuspended . The method automatically restores all the corrupt passive copies from the backup created in the last backup session (and the corresponding restore chain). After the data is restored, the copies are synchronized with the active copy, provided that the Resume database replication option is selected.
Restore to the latest state
This method is used to restore a corrupt database to the latest possible point in time. Data Protector restores the database from the backup created in the last backup session (and the corresponding
restore chain). For details, see Restore chain, on page 255 .
Once the files are restored, all the logs (not only those restored from the backup, but also any existing logs) are replayed to the database file.
NOTE:
DAG environments:
When a passive copy is restored, Microsoft Exchange Server ensures that the logs are replayed to the database file in accordance with the
ReplayLagTime
parameter setting.
Restore to a point in time
This method is used to restore a database to a specific point in time.
NOTE:
Standard restore:
When you restore a standalone database or active copy, the existing .log
and .chk
files are renamed (a .keep
extension is added to their names). This feature is useful when you restore files without performing database recovery. It enables you to apply additional logs to the database file; just delete the
.keep
extension of the log files that you also want to be applied and start a database recovery manually. In this way, you can fine-tune the point in time the database is restored to.
When you restore a passive copy, the existing files are deleted.
Once the files are restored, the logs are replayed to the database file ( .edb
) if the Perform database
recovery option is selected.
Data Protector (10.00) Page 252 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
NOTE:
DAG environments:
l
When a passive copy is restored, Microsoft Exchange Server ensures that the logs are replayed to the database file in accordance with the ReplayLagTime parameter setting.
l
For passive copies that are not restored, a full reseed is required once the restore session completes.
Restore to a new mailbox database
This method is used to restore data to a different database, either because the original database no longer exists or in order to move the data elsewhere.
Using it, you can restore data also to a Microsoft Exchange Server recovery database.
NOTE:
Instant recovery:
This option is not available for replica types whose data can only be restored to the original storage volumes.
Restore files to a temporary location
Using this method, you can restore database files to a location of your choice.
l
When you restore from a Differential or Incremental backup session, you can restore the complete restore chain or only the files ( .log
) backed up in the selected session.
l
When you restore data from a Full backup session, you have an option to restore only the database file ( .edb
).
NOTE:
Instant recovery:
This option is not available for replica types whose data can only be restored to the original storage volumes.
Restore destination
Backup data can be restored: l to an existing database (standalone database, active copy, passive copy), l to a new database, l to a temporary location.
Restoring to a standalone database
Restore to the original standalone database (standalone environment) progresses as follows:
1. The database is dismounted.
2. Backup data is restored.
Data Protector (10.00) Page 253 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
3. Optionally, the newly-restored logs (and pre-existing ones if you are performing the Restore to the
latest state method) are replayed to the database file
.edb
and the database is mounted.
To restore to the original standalone database, use one of the following restore methods: l
Restore to the latest state
l
Restore to a point in time
Restoring to an active copy
Restore to the active copy (DAG environment) progresses as follows:
1. The database is dismounted.
2. All replications are suspended.
3. Backup data is restored.
4. Optionally, the newly-restored logs (and pre-existing ones if you are performing the Restore to the
latest state method) are replayed to the database file .edb
and the database is mounted.
To restore to the active copy, use one of the following restore methods: l
Restore to the latest state
l
Restore to a point in time
Restoring to a passive copy
Restore to a passive copy (DAG environment) progresses as follows:
1. The replication is suspended.
2. Backup data is restored.
3. Optionally, the replication with the active copy is resumed.
To restore to a passive copy, use one of the following restore methods: l
Restore all passive copies with failed status
l
Restore to the latest state
l
Restore to a point in time
Restoring data to a new database
Restore to a new database progresses as follows:
1. A new mailbox database is created.
2. Backup data is restored to the new database.
NOTE:
If you restore to a recovery database, first the backup data is restored and then a recovery database is created.
To restore data to a new mailbox database or recovery database, use the Restore to a new mailbox
database restore method.
Data Protector (10.00) Page 254 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restoring data to a temporary location
You can restore the database file ( .edb
and/or .log
and/or .chk
) to a client and directory of your choice. Select the Restore files to a temporary location restore method.
Instant recovery in DAG environments
When you back up a database in a DAG environment, you can decide whether to back up all its copies or only a single copy. If all copies are backed up, during instant recovery, all copies can be restored, as each copy has its own replica storage volumes to be restored from. If only a single copy is backed up, note the following: l
In most cases, only one database copy can be restored from the backup of a single database copy.
Some replica types are directly connected with the source (dependent replica types) while others are
independent, allowing the data to be restored to a different location. With the latter, you can restore either the original or a different database copy.
NOTE:
The following replica types are independent: o
HPE P9000 XP Disk Array Family (split mirror replica type in the VSS compliant mode) o
HPE P6000 EVA Disk Array Family (snapclone replica type)
For dependent replica types, Data Protector automatically grays out those clients in the Target
Nodes option whose database copies were not backed up, because these copies cannot be restored.
l
The only replica type that enables you to restore multiple database copies from the backup of a single database copy is snapclone (only with the HPE P6000 EVA Disk Array Family). However, you must also ensure that both the Restore using HPEP6000 EVA SMI-S and Copy replica data
to the source volumes instant recovery options are selected, in which case data from the replica storage volumes is sequentially copied to multiple locations, restoring one database copy after the other.
Restore chain
By default, when you select a Differential or Incremental backup session for restore, Data Protector restores not only the logs ( .log
) backed up in the selected session but also files backed up in preceding sessions (restore chain): l
If a Differential backup session is selected, Data Protector restores:
1. The .edb
file and .log
files backed up in the Full or Copy backup session.
2. The .log
files backed up in the selected Differential backup session.
l
If an Incremental backup session is selected, Data Protector restores:
1. The .edb
file and
.log
files backed up in the Full or Copy backup session.
2. The .log
files backed up in all subsequent Incremental backup sessions, up to the selected
Incremental backup session.
Data Protector (10.00) Page 255 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration l
If a Full or Copy backup session is selected, Data Protector restores the .edb
file and .log
files backed up in the selected session.
NOTE:
l
If the Restore to the latest state method is used, .log
files from the Full or Copy backup session are not restored.
l
The only method that enables you to restore only
.log
files backed up in the selected
Incremental or Differential session is Restore to a temporary location.
Restore chain during instant recovery
During an instant recovery session, you first select which: l
Full (ZDB-to-disk or ZDB-to-disk+tape) session or l
Copy (ZDB-to-disk or ZDB-to-disk+tape) session to use for instant recovery. From database specific options you then specify whether additional logs should also be restored, by selecting a subsequent Incremental or Differential session in the Restore
additional logs until option.
In an instant recovery session Data Protector restores:
1. The .edb
file and .log
files backed up in the selected Full or Copy backup session.
2. The .log
files backed up in the selected Differential backup session or in all subsequent
Incremental backup sessions, up to the selected Incremental backup session.
Restore parallelism
If device concurrency allows, database copies are restored in parallel, except in the following cases: l
If database copies were backed up from the same client, but are now restored to different clients.
l
If backup data of the same database copy is used as a restore source for multiple database copies.
Finding information for restore
You can retrieve information about backup sessions (such as information on the backup type and media used, and the messages reported during the backup) from the Data Protector IDB.
To retrieve information, use the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Internal Database.
2. In the Scoping pane, expand Objects or Sessions.
If you expand Objects, backup objects are sorted according to the Microsoft Exchange Server databases for which they were created.
NOTE:
The backup object name contains the database GUID. To find out which GUID belongs to
Data Protector (10.00) Page 256 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration which database, see the database object /DB_GUID/DB_Name .
For example, the database object for the database
DB1
with the GUID
08bca794-c544-4e27-87e8-533fb81fd517 is:
/08bca794-c544-4e27-87e8-533fb81fd517/DB1
If you expand Sessions, backup objects are sorted according to the sessions in which they were created. For example, backup objects created in the session
2013/02/7–7
are listed under
2013/02/7–7 .
To view details on a backup object, right-click the backup object and click Properties.
TIP:
To view the messages reported during the session, click the Messages tab.
Using the Data Protector CLI
1. Log in to the Cell Manager or to any Microsoft Exchange Server client with the Data Protector
User Interface component installed under a user account that is configured as described in
Configuring user accounts, on page 234
.
2. Get a list of Microsoft Exchange Server backup objects created in a backup session: omnidb -session SessionID
3. Get details on a backup object: omnidb -e2010 BackupObjectName -session SessionID -catalog
Here is one example of a backup object name: devy.company.com:/08bca794-c544-4e27-87e8-533fb81fd517/DB1
For details, see the omnidb man page or the HPE Data Protector Command Line Interface Reference.
Standard restore
Standard restore is restore from backup data residing on Data Protector media (for example, a tape).
Such data is created in ZDB-to-disk+tape and ZDB-to-tape sessions.
You can restore multiple Microsoft Exchange Server databases in the same standard restore session,
specifying a different restore method for each database. For details, see Restore methods, on page
.
To perform a standard restore, use the Data Protector GUI or CLI.
Restoring using the Data Protector GUI
1. In the Context List, click Restore.
2. In the Scoping Pane, expand MS Exchange 2010+ Server, expand the DAG virtual system or standalone Microsoft Exchange Server system and click MS Exchange 2010+ Server.
3. In the Source page, Data Protector displays all Microsoft Exchange Server databases backed up from the selected DAG or standalone environment.
Select the Microsoft Exchange Server databases to restore.
Data Protector (10.00) Page 257 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
When you select a database, the Properties for Database dialog box is displayed automatically.
Specify a restore method and click OK. For databases that are part of a DAG, the default restore method is Repair all passive copies with failed status. For standalone databases, the default is Restore to the latest state.
Selecting databases for restore
To change the restore method, right-click the database and click Properties.
Data Protector (10.00) Page 258 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Repair all passive copies with failed status
For details, see Repair all passive copies with failed status, on page 275
.
Data Protector (10.00) Page 259 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore to the latest state
For details, see Restore to the latest state, on page 252 .
Data Protector (10.00) Page 260 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore to a point in time
For details, see Restore to a point in time, on page 252 .
Data Protector (10.00) Page 261 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore to a recovery database
For details, see Restore to a new mailbox database, on page 253
.
Data Protector (10.00) Page 262 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore files to a temporary location
For details, see Restore files to a temporary location, on page 253 .
4. In the Options page, specify the Data Protector Microsoft Exchange Server 2010 integration
restore options. For details, see General restore options, on page 279 .
Data Protector (10.00) Page 263 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Restore options
5. In the Devices page, select the devices to use for restore.
For details on how to select devices to be used for restore, see the HPE Data Protector Help index: “restore, selecting devices for”.
6. Click Restore.
7. In the Start Restore Session dialog box, click Next.
8. Specify Report level and Network load.
NOTE:
Select Display statistical information to view the restore profile messages in the session output.
9. Click Finish to start the restore.
The statistics of the restore session, along with the message Session completed successfully is displayed at the end of the session output.
Restoring using the Data Protector CLI
1. Log in to the Cell Manager or to any Microsoft Exchange Server client with the User Interface
.
2. Execute the following: omnir -e2010
-barhost
ClientName
[ VSS_EXCHANGE_SPECIFIC_OPTIONS]
Database [Database …]
[-user
User:Domain]
Data Protector (10.00) Page 264 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
[ GENERAL_OPTIONS]
Database
{-db_name
SourceDatabaseName | -db_guid SourceDatabaseGUID }
[-source
SourceClientName]
{-repair | -latest | -pit | -new | -temp}
E2010_METHOD_OPTIONS E2010_REPAIR_
METHOD_OPTIONS
[–no_resume_replication]
E2010_LATEST_METHOD_OPTIONS
[-node TargetNode … | -all]
[–no_resume_replication]
[–no_recover]
[–no_mount]
E2010_PIT_METHOD_OPTIONS
-session
BackupID
[-node TargetNode … | -all]
[–no_resume_replication]
[–no_recover]
[–no_mount]
E2010_NEW_METHOD_OPTIONS
-session
BackupID
-client
TargetClientName
-location
TargetDatabasePath
-name
TargetDatabaseName
[-recoverydb]
[–no_recover]
[–no_mount]
E2010_TEMP_METHOD_OPTIONS
-session
BackupID
-client
TargetClientName
-location
TargetDatabasePath
[–no_chain]
[–edb_only]
[–no_recover]
For a brief description of the options, see Restore options, on page 275
. For details, see the omnir man page or the HPE Data Protector Command Line Interface Reference.
NOTE:
A backup ID is a point in time. All objects (backup data) created in a backup session have the same backup ID, which is the same as the session ID of the backup session.
Mirrored objects and objects created in an object copy session have the same backup ID as objects created in the original backup session. Suppose the media set created in the original backup session no longer exists, but the media set created in an object copy
Data Protector (10.00) Page 265 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration session still exists. To restore the objects, you must specify the session ID of the original backup session (that is, the backup ID) and not the session ID of the object copy session.
The omnir syntax does not let you specify from which object copy to restore if several copies of the same object exist. This is only possible using the Data Protector GUI by setting the media allocation priority list.
Example (Restore method – repair)
DAG environment
To restore all corrupt passive copies of the database DB1 , which was backed up from a DAG whose virtual system name was dag0.company.com
, and to ensure that the integration agent ( e2010_ bar.exe
) is started on the client exchange2.company.com
, execute: omnir -e2010 -barhost exchange2.company.com –db_name DB1 –source dag0.company.com repair
Example (Restore method – latest)
Standalone environment
To restore the corrupt standalone database
DB1
, which resides on the client exchange1.company.com
, to the latest possible point in time, and to ensure that the integration agent ( e2010_bar.exe
) is started on the client exchange2.company.com
, execute: omnir -e2010 -barhost exchange2.company.com –db_name DB1 –source exchange1.company.com –latest
DAG environment
Suppose you want to restore the active copy of the database DB1 , which resides on the client exchange1.company.com
, and the passive copies of the database that reside on the clients exchange2.company.com
and exchange3.company.com
. Suppose the database
DB1
is part of a DAG whose virtual system name is dag0.company.com
, and that you want the integration agent ( e2010_ bar.exe
) to be started on the client exchange2.company.com
. Execute the following command: omnir -e2010 -barhost exchange2.company.com –db_name DB1 –source dag0.company.com – latest –node exchange1.company.com –node exchange2.company.com –node exchange3.company.com
Example (Restore method – pit)
Standalone environment
Suppose you want to restore the corrupt standalone database DB1 , which resides on the client exchange1.company.com
, using the backup data created in the session
2013/5/14-1
. Suppose you want the integration agent ( e2010_bar.exe
) to be started on the client exchange1.company.com
.
Execute the following command: omnir -e2010 -barhost exchange1.company.com –db_name DB1 -pit –session 2013/5/14-1
NOTE:
The
–source
option is not specified, in which case Data Protector assumes that the database was backed up from the client specified with the
–barhost
option.
Data Protector (10.00) Page 266 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Example (Restore method – new)
DAG environment
Suppose you want to restore the backup of the database
DB1
to a recovery database that should be created on the client exchange2.company.com
and named Recovery1 , with the files in the
C:\Recovery1Folder directory. Suppose the database DB1 was backed up in the session 2013/5/14-
1 from a DAG whose virtual system name was dag0.company.com
. To also ensure that the integration agent ( e2010_bar.exe
) is started on the client exchange1.company.com
, execute the following command: omnir -e2010 -barhost exchange1.company.com –db_name DB1 –source dag0.company.com – new –session 2013/5/14-1 –client exchange2.company.com -location C:\Recovery1Folder
–name Recovery1 –recoverydb
Example (Restore method – temp)
Standalone environment
Suppose you want to restore the transaction logs of the database DB1 , which resides on the client exchange2.company.com
. The logs were backed up in the Incremental backup session 2013/5/14-1 .
To restore the logs to the client exchange2.company.com
to the directory
C:\DB1TransactionLogFolder without performing database recovery, and to ensure that the integration agent ( e2010_bar.exe
) is started on the client exchange1.company.com
, execute the following command: omnir -e2010 -barhost exchange1.company.com –db_name DB1 –source exchange2.company.com –temp –session 2013/5/14-1 –client exchange2.company.com – location C:\DB1TransactionLogFolder -no_chain –no_recover
Restoring using another device
You can restore using a device other than that used for a backup. For details, see the HPE Data
Protector Help index: “restore, selecting devices for”.
Instant recovery
To be able to perform an instant recovery session, you need backup data that is stored on replica storage volumes. Such backup data is created in ZDB-to-disk and ZDB-to-disk+tape sessions.
You can restore multiple Microsoft Exchange Server databases in the same instant recovery session,
specifying a different restore method for each database. For details, see Restore methods, on page
.
To start an instant recovery session, use the Data Protector GUI or CLI.
Performing instant recovery using the Data Protector GUI
To perform an instant recovery session:
1. In the Context List, click Instant Recovery.
2. Expand Exchange 2010+ Server and select which ZDB-to-disk or ZDB-to-disk+tape session to
Data Protector (10.00) Page 267 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration use for instant recovery. The sessions are sorted according to the backup specifications used.
3. In the Source page, select which Microsoft Exchange Server databases to restore.
When you select a database, the Properties for Database dialog box is displayed automatically.
Specify a restore method and click OK. For databases that are part of a DAG, the default restore method is Repair all passive copies with failed status. For standalone databases, the default is Restore to the latest state. To change the restore method, right-click the database and click
Properties.
For details on Configuration check mode, press F1.
Selecting databases for instant recovery
Restore to the latest state
Data Protector (10.00) Page 268 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For details, see Restore to a point in time, on page 276 .
Restore to a point in time
Data Protector (10.00) Page 269 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For details, see Restore to a point in time, on page 276 .
Restore to a new mailbox database
Data Protector (10.00) Page 270 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For details, see Restore to a new mailbox database, on page 253
.
Restore files to a temporary location
Data Protector (10.00) Page 271 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
For details, see Restore files to a temporary location, on page 253 .
4. In the Options page, specify ZDB-specific options. For details, press F1.
5. In the Advanced page, specify the Data Protector Microsoft Exchange Server 2010 integration
instant recovery options. For details, see
General restore options, on page 279
Instant recovery – advanced
Data Protector (10.00) Page 272 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
6. Click Restore.
Performing instant recovery using the Data Protector CLI
1. Log in to the Data Protector Cell Manager or to any Microsoft Exchange Server client under a user
account as described in Configuring user accounts, on page 234
.
2. Execute the following: omnir -e2010
-barhost ClientName
-instant_restore
[
VSS_INSTANT_RECOVERY_OPTIONS]
[
VSS_EXCHANGE_SPECIFIC_OPTIONS]
Database [Database …]
[-user
User:Domain]
[ GENERAL_OPTIONS]
Database
{-db_name SourceDatabaseName | -db_guid SourceDatabaseGUID}
[-source
SourceClientName]
{-repair | -latest | -pit | -new | -temp}
E2010_METHOD_OPTIONS E2010_REPAIR_
METHOD_OPTIONS
[–no_resume_replication]
Data Protector (10.00) Page 273 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
E2010_LATEST_METHOD_OPTIONS
[-node TargetNode … | -all]
[–no_resume_replication]
[–no_recover]
[–no_mount]
[E2010_IR_SPECIFIC_OPTIONS] E2010_PIT_METHOD_OPTIONS
-session
SessionID
[-node TargetNode … | -all]
[–no_resume_replication]
[–no_recover]
[–no_mount]
[E2010_IR_SPECIFIC_OPTIONS] E2010_NEW_METHOD_OPTIONS
-session
SessionID
-client
TargetClientName
-location
TargetDatabasePath
-name
TargetDatabaseName
[-recoverydb]
[–no_recover]
[–no_mount]
[E2010_IR_SPECIFIC_OPTIONS] E2010_TEMP_METHOD_OPTIONS
-session
SessionID
-client
TargetClientName
-location
TargetDatabasePath
[–no_chain]
[–edb_only]
[–no_recover]
[E2010_IR_SPECIFIC_OPTIONS] E2010_IR_SPECIFIC_OPTIONS
[-from_session
SessionID]
For brief description of the options, see the section Restore options, on the next page
.
For details on the options, see the HPE Data Protector Command Line Interface Reference or the omnir
man page.
Example (Restore method – latest)
Standalone environment
Suppose you want to restore the corrupt standalone database DB1 , which resides on the client exchange1.company.com
. To ensure the integration agent ( e2010_bar.exe
) is started on the client exchange1.company.com
, and that the database is restored to the latest state, execute: omnir -e2010 -barhost exchange1.company.com –instant_restore –copy_back –db_name
DB1 -latest
Example (Restore method – temp)
DAG environment
Suppose you want to restore the database DB1 , which was part of a DAG whose DAG virtual system
(host) name was dag0.company.com
. The database was backed up in the session 2013/5/14-1 . To
Data Protector (10.00) Page 274 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration restore the database to a temporary location on the client exchange1.company.com
to the directory
C:\BackupDatabase , and to ensure that the integration agent ( e2010_bar.exe
) is started on the client exchange1.company.com
, execute: omnir -e2010 -barhost exchange1.company.com –instant_restore –copy_back –db_name
DB1 –source dag0.company.com –temp –session 2013/5/14-1 –client exchange1.company.com –location C:\BackupDatabase
Restore options
Repair all passive copies with failed status
Option in the GUI / CLI Description
/
Resume database replication
-no_resume_replication
Available in DAG environments. Resumes the replication between the active and passive copies after the copies are restored.
Note that the CLI option -no_resume_replication has the opposite meaning. If it is specified, the replication in not resumed.
Restore additional logs until
-session
Target nodes
This is an instant recovery-specific option.
Not available.
Not available.
The clients (that is, copies) that have the status
Failed
or
FailedAndSuspended are automatically selected.
Restore to the latest state
Option in the GUI / CLI Description
Select for restore
Restore additional logs until
-session
Perform database recovery /
-no_recover
Mount database /
Specifies whether the database should be restored.
This is an instant recovery-specific option.
Not available.
Available when restoring a standalone database (standalone environment) or an active copy (DAG environment). Applies the logs to the database file ( .edb
) after the restore completes.
Note that the CLI option -no_recover has the opposite meaning. If it is specified, the database recovery in not performed.
Available when restoring a standalone database (standalone environment) or an active copy (DAG environment). Mounts the
Data Protector (10.00) Page 275 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Option in the GUI / CLI
-no_mount
/
Resume database replication
-no_resume_replication
Available when restoring passive copies (DAG environment).
Resumes the replication between the active and passive copies after the copies are restored.
Note that the CLI option -no_resume_replication has the opposite meaning. If it is specified, the replication is not resumed.
Target nodes
-node | –all
Restore to a point in time
Available only in DAG environments. Specifies which clients
(that is, database copies) to restore.
Option in the GUI / CLI
Select for restore
Description
database after the database recovery completes. This option is available only if Perform database recovery is selected.
Note that the CLI option
-no_mount
has the opposite meaning.
If it is specified, the database is not mounted.
Backup version /
-session
Last backup version
Restore additional logs until
-session
Description
See the description in Restore to the latest state, on the previous page
.
This is a standard restore-specific option.
It specifies from which backup data to restore. Select a backup
ID.
If a Differential backup session is selected, the .log
files backed up in the selected Differential backup session are restored.
If an Incremental backup session is selected, the
.log
files backed up in all subsequent Incremental backup sessions, up to the selected Incremental backup session, are restored.
This is a standard restore-specific option.
It shows the session in which the database was last backed up.
This is an instant recovery-specific option.
If a Differential backup session is selected, the .log
files backed up in the selected Differential backup session are restored.
If an Incremental backup session is selected, the .log
files backed up in all subsequent Incremental backup sessions, up to the selected Incremental backup session, are restored.
Data Protector (10.00) Page 276 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Option in the GUI / CLI
Perform database recovery /
-no_recover
Description
See the description in Restore to the latest state, on page 275 .
Mount database /
-no_mount
/
Resume database replication
-no_resume_replication
Target nodes /
-node | –all
See the description in Restore to the latest state, on page 275 .
The node (client) hosting the active copy is automatically selected for restore.
Restore to a new mailbox database
Option in the GUI / CLI
Select for restore
Restore additional logs until
-session
Description
See the description in Restore to the latest state, on page 275 .
This is an instant recovery-specific option.
See the description in Restore to a point in time, on the previous page .
Specifies the client to restore to.
Target client /
-client
Restore into location /
-location
Specifies the directory to restore to (standard restore) or the directory to mount the replica storage volumes to (instant recovery).
Specifies the name to be used for the new database. If another database with the same name already exists, the restore fails.
Database name /
-name
Restore into Recovery
database /
-recoverydb
Backup version /
-session
Restores the data to a Microsoft Exchange Server recovery database.
Although multiple recovery databases can exist in parallel, only one recovery database can be mounted to the Microsoft
Exchange Server at a time.
See the description in Restore to a point in time, on the previous page .
Data Protector (10.00) Page 277 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Option in the GUI / CLI
Last backup version
Perform database recovery /
-no_recover
Mount database /
-no_mount
Target nodes
Description
Not available.
Restore to the latest state, on page 275
Restore files to a temporary location
Option in the GUI / CLI Description
Select for restore
Restore additional logs until
-session
Restore chain
See the description in Restore to the latest state, on page 275
This is an instant recovery-specific option.
See the description in Restore to a point in time, on page 276
.
If this option is set to Restore only this backup, only files backed up in the selected session are restored.
If this option is set to Full restore (full, incr, diff backups), the complete chain is restored.
See the description in Restore to a new mailbox database, on the previous page .
Target client /
-client
Restore into location /
-location
Backup version /
-session
Last backup version
Restore database files only /
-edb_only
Perform database recovery /
-no_recover
Target nodes
See the description in Restore to a point in time, on page 276
Restores only the database files ( checkpoint files (
Not available.
.chk
.edb
). The logs (
) are not restored.
.log
) and
.
Restore to the latest state, on page 275 .
Data Protector (10.00) Page 278 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
General restore options
Option in the GUI / CLI
Startup client /
-barhost
Username
Group/Domain name /
-user
Perform consistency check /
[-exch_check
[ -exch_throttle Value ] |
-exch_checklogs]
Description
Specifies the client on which the integration agent ( e2010_ bar.exe
) should be started. If the DAG virtual client (host) is selected, the integration agent is started on the currently active
node. To find out which node is currently active, see Restore, on page 251 .
Default: The same client that was specified for the backup session. If the DAG virtual client was specified, this client is now selected. However, note that the integration agent may not be started on the same physical node as during the backup session; it depends which node is currently active.
Specifies which Windows domain user account to use for the restore session. Ensure that the user is configured as described
in Configuring user accounts, on page 234
.
If these options are not specified, the restore session is started under the user account under which the Data Protector Inet service is running.
If this option is selected, Microsoft Exchange Server checks the consistency of a database's backup data. If this option is not selected, the session finishes earlier, but the backup data consistency is not guaranteed.
The check is performed at the target locationon the source storage volumes after the backup data is restored. You do not need to perform the consistency check if it was already performed at the time of backup.
Default: not selected
If the Check log files only option is selected, only the log file backup data is checked, which is enough for Microsoft
Exchange Server to guarantee data consistency.
Default: not selected
By default, the consistency check is I/O intensive, which can negatively affect disk performance. The Throttle check for 1
second option throttles down the consistency check of the database file .edb
to lessen impact on the disk performance.
Specify after how many input/output operations the check should stop for one second.
This option is not available if only the log files are checked.
Default: not selected
Data Protector (10.00) Page 279 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
TIP:
To find out which Microsoft Exchange Server node is currently active, connect to one of the nodes and execute: cluster group
Example
C:\Users\administrator.E2010BETA>cluster group
Listing status for all available resource groups:
Group Node Status
-------------------- --------------- ------
Available Storage spade Offline
Cluster Group club Online
The currently active node has the status Online . In the example, this is club .
Monitoring sessions
You can monitor currently running sessions in the Data Protector GUI. When you run a backup or restore session, a monitor window shows the progress of the session. Closing the GUI does not affect the session.
You can also monitor sessions from any Data Protector client with the User Interface component installed, using the Monitor context.
To monitor a session, see the HPE Data Protector Help index: “viewing currently running sessions”.
Troubleshooting
This section lists general checks and verifications, plus problems you might encounter when using the
Data Protector Microsoft Exchange Server 2010 integration.
Because the Data Protector Microsoft Exchange Server 2010 integration is based on the Data
Protector Microsoft Volume Shadow Copy Service integration, also see troubleshooting information in the HPE Data Protector Integration Guide for Microsoft Volume Shadow Copy Service.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
Before you begin
l
Ensure that the latest official Data Protector patches are installed. On how to verify this, see the
HPE Data Protector Help index: “patches”.
l
See the HPE Data Protector Product Announcements, Software Notes, and References for general
Data Protector limitations, as well as recognized issues and workarounds.
l
See https://softwaresupport.hpe.com/manuals for an up-to-date list of supported versions, platforms, and other information.
Data Protector (10.00) Page 280 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Checks and verifications
If your browsing, backup, or restore failed: l
Examine system errors reported in the debug.log
file.
l
Check if you can do a filesystem backup and restore on the problematic client. For information, see the HPE Data Protector Help.
Problems
Problem
It takes a long time to display Microsoft Exchange Server topology in the Data Protector GUI
When you open the Data Protector GUI and try to display the source page, either in the Backup or
Restore context, you must wait a long time.
This may happen if there is an unresponsive system in the same domain (for example, a system that is shut down). The problem occurs even if the unresponsive system is not part of your backup environment. This is due to Microsoft Exchange Server problems with execution of Microsoft
Exchange Server Shell commands.
Action
Remove the system from the domain or fix the problem.
Problem
A database backup cannot be performed
When you start a backup session for a database, the database is not backed up, appearing to be locked by other session, though there are no other backup sessions currently running. A message similar to the following is displayed:
[Minor] From: [email protected] "MS Exchange
Server" Time: 1/17/2013 3:07:13 PM
[170:313] One or more copies of database DEMAR are already being backed up in a different session.
This may happen if the integration agent ( e2010_bar.exe
) was terminated by force while a previous backup session was in progress, either because the Microsoft Exchange Server system was restarted or for some other reason, so the lock remains.
Action
Execute the following command: omnidbutil –free_cell_resources
NOTE:
This command line removes all existing locks, so ensure that none of the existing locks is still needed.
Data Protector (10.00) Page 281 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
Problem
Restore fails
When you try to restore a database, the session fails.
This may happen if a database has been restored before (probably unsuccessfully), and during that previous restore session, the Microsoft Exchange Server created an .env
file in the database directory.
This file now prevents the database from being restored again.
Action
Delete the .env
file and start a new restore session.
Problem
Restore from an object copy fails in a DAG environment
When restoring a database from a media set created in an object copy session, as the media set created in the original backup session no longer exists, the session fails with an error similar to the following:
[Critical] From: [email protected] "MS Exchange 2010 Server"
Time: 28/02/2013 16:08:12 No mailbox database copy can be selected for restore/instant recovery.
Action
1. Verify that the media set created in the object copy session still exists.
2. On all Microsoft Exchange Server system nodes, set the environment variable OB2BARHOSTNAME to the name of the DAG virtual system and restart the Data Protector Inet service.
3. Start a new restore session.
Problem
Restore to the latest state fails
When you try to restore a database all of whose log files were lost, using the restore method Restore
to the latest state with the Perform database recovery option selected, the database recovery fails.
This may happen if a database is restored from a Full backup (that is, the restore chain consists of only the Full backup session). Since in the Restore to the latest state session, only the
.edb
file is
there are no logs to be applied to the database file, and the database recovery fails.
Action
Problem
After instant recovery to a point in time, passive copies remain in the Failed state
When you start an instant recovery session in a DAG environment to restore active and passive copies of the same database, using the restore method Restore to a point in time, the data is successfully
Data Protector (10.00) Page 282 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration restored, but the synchronization between the active copy and passive copies fails, leaving the passive copies in the Failed state.
This problem occurs if, after the data is restored, the passive copy has extra log files, which are not present at the active copy side, and so synchronization cannot be established. This can happen if, during a Full backup session, multiple copies of a database are selected for backup. Data Protector first performs a Full backup of the passive copy that has the fewest logs applied to the database file, and then performs a Copy backup of all the remaining copies, with the active copy being backed up last. While the backup session is in progress, a new log may be created at the active copy side, so when the active copy is backed up, the newly created log is also backed up. If, further on in time, a failover occurs (one of the passive copies becomes the active copy) and you perform a Restore to a
point in time instant recovery, each copy is restored from its own replica storage volumes. This results in the active copy (which was passive at the time of backup) having fewer logs than the passive copy (which was active at the time of backup). Consequently, synchronization cannot be established.
Action
Perform a full reseed for all Failed passive copies.
Problem
In a DAG, a copy-back instant recovery fails when restoring a non-original database copy
Suppose you backed up a database copy by creating a snapclone replica (HPE P6000 EVA Disk Array
Family). Using the copy-back method, this replica can be used to restore the original database copy and/or the related database copies that reside on different Microsoft Exchange Server systems in the
DAG. If the size of the source storage volumes on those Microsoft Exchange Server systems differs from the size of the source storage volumes that were backed up, the instant recovery session fails.
If the Retain source for forensics option is selected, a message similar to the following is displayed:
[Warning] From: [email protected] "SMISA" Time:
1/17/2013 2:51:08 PM
[236:8001] This pair of source and target storage volumes are not the same size.
Source storage volume : 50014380025B4860\\Virtual Disks\VSS\
FizzyDizzy\DAG\dizzy\dizzy-DB1-data\ACTIVE
Source size : 4 GB
Target storage volume : 50014380025B4860\\Virtual Disks\VSS\
FizzyDizzy\DAG\fizzy\hpVSS-LUN-06Jul10 02.23.27\ACTIVE
Target size : 3 GB
[Major] From: [email protected] "MS Exchange
Server" Time: 1/17/2013 3:11:16 PM
The system failed to refresh symbolic links in kernel object namespace.
If the Retain source for forensics option is not selected, a message similar to the following is displayed:
[Major] From: [email protected] "SMISA" Time:
1/17/2013 2:51:08 PM
[236:8001] This pair of source and target storage volumes are not the same size.
Source storage volume : 50014380025B4860\\Virtual Disks\VSS\
Data Protector (10.00) Page 283 of 413
Zero Downtime Backup Integration Guide
Chapter 4: Data Protector Microsoft Exchange Server 2010+ ZDB integration
FizzyDizzy\DAG\dizzy\dizzy-DB1-data\ACTIVE
Source size : 4 GB
Target storage volume : 50014380025B4860\\Virtual Disks\VSS\
FizzyDizzy\DAG\fizzy\hpVSS-LUN-06Jul10 02.23.27\ACTIVE
Target size : 3 GB
Action
Ensure that storage volumes on different Microsoft Exchange Server systems are the same size.
Data Protector (10.00) Page 284 of 413
Chapter 5: Data Protector Microsoft
SharePoint Server Server VSS based solution
Introduction
This chapter explains how to configure and use the Data Protector Microsoft SharePoint Server VSS based solution (VSS based solution). In reality, the solution is based on the Data Protector Microsoft Volume
Shadow Copy Service integration (VSS integration). For details on the VSS integration, see the HPE Data
Protector Integration Guide for Microsoft Volume Shadow Copy Service.
The chapter describes concepts and methods you need to understand to back up and restore Microsoft
SharePoint Server 2010 and Microsoft SharePoint Server 2013 data that is stored in Microsoft SQL Server databases. For example: l
The configuration database (
SharePoint_Config
) l
Content databases ( SharePoint_AdminContent_Label , WSS_Content_Label ,...) l
SharePoint Service Applications databases ( SSA_DB ) (Microsoft SharePoint Server 2010/2013)
In addition, you can also back up and restore Microsoft SharePoint Server search index files.
From now on, both Microsoft SharePoint Server versions are called Microsoft SharePoint Server, unless the differences are pointed out.
Backup
Microsoft SharePoint Server data that is stored in Microsoft SQL Server databases is backed up using one of the following Microsoft SQL Server VSS writers: l
MSDE writer (for Microsoft SQL Server 2000 databases) l
SqlServerWriter (for Microsoft SQL Server 2008 databases)
Microsoft SharePoint Server 2010 search index files are backed up using the following VSS writers: l
OSearch14 VSS writer l
SPSearch4 VSS writer
Microsoft SharePoint Server 2013 search index files are backed up using following VSS writers: l
OSearch15 VSS writer
Microsoft FAST Search Server 2010 search index files are backed up: l using the Data Protector Disk Agent (in case of the standard filesystem backup with VSS enabled) l using the Data Protector VSS integration (in case of the ZDB filesystem backup)
You can create and run backup specifications using the Data Protector PowerShell command which is
described in Backup, on page 289
.
Data Protector (10.00) Page 285 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Limitations
l
The only supported way to start backup sessions is using the Data Protector PowerShell command.
Starting the backup sessions using the Data Protector GUI or CLI is not supported.
l
Microsoft SharePoint Server 2010: With VSS based solution, the FAST Search index files can also be backed up incrementally when using the Data Protector Disk Agent. For all other Microsoft
SharePoint Server data only Full backup type is supported.
Restore
Restore can be started using the Data Protector GUI or CLI as described in Restore, on page 300
.
Installation and configuration
ZDB prerequisites
If you plan to run ZDB and instant recovery (IR) sessions, ensure that the SPSearch and OSearch index files of each SSP or SSA, and the FAST Search index files reside on a disk array.
Microsoft Office SharePoint Server 2007
The default location for the SPSearch index files is:
C:\Program Files\Microsoft Office Servers\12.0\Data\Applications
The default location for the OSearch index files is:
C:\Program Files\Microsoft Office Servers\12.0\Data\Office Server\Applications
To move the index files to the disk array:
1. Open the Command Prompt and change the directory to:
C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\12\BIN>
2. To move the SPSearch index files, execute: stsadm -o spsearch -indexlocation PathToNewLocation
3. To move the OSearch index files, execute: stsadm -o editssp –title SSPname -indexlocation PathToNewLocation
Microsoft SharePoint Server 2010
The default location for the SPSearch index files is:
C:\Program Files\Microsoft Office Servers\14.0\Data\Applications
The default location for the OSearch index files is:
C:\Program Files\Microsoft Office Servers\14.0\Data\Office Server\Applications
Data Protector (10.00) Page 286 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
To move the index files to the disk array:
1. Open the Command Prompt and change the directory to:
C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\14\BIN>
2. To move the SPSearch index files, execute: stsadm -o spsearch -indexlocation PathToNewLocation
3. To move the OSearch index files use the Central Administration (modify farm topology).
The
FASTSearch
home folder must be installed to the disk array during the FAST Search Server 2010 system installation. In case of a multiple FAST Search Server system farm, ensure that the
FASTSearch home folders on all the systems have the same path (drive and path name).
Licensing
The Data Protector VSS based solution requires one online-extension license per each Microsoft
SharePoint Server client participating in the backup and restore process. This means one onlineextension license for each system on which the Data Protector MS Volume Shadow Copy
Integration component is installed.
Installing the integration
For details on how to install a Data Protector cell, see the HPE Data Protector Installation Guide.
To be able to back up Microsoft SharePoint Server objects, install the following installation packages and Data Protector components: l
Windows PowerShell 2.0 or later and the Data Protector
User Interface
component on the
Microsoft SharePoint Server system on which you plan to execute the Data Protector commands and on which you install the Data Protector MS Volume Shadow Copy Integration component.
See the next bullet.
If not already available on your Windows system, you can download Windows PowerShell from http://www.microsoft.com/windowsserver2003/technologies/management/powershell/ default.mspx
.
l
The Data Protector
MS Volume Shadow Copy Integration
component on the Microsoft SQL
Server system and the Microsoft SharePoint Server systems that have at least one of the following services enabled:
Microsoft Office SharePoint Server 2007:
o
Windows SharePoint Services Database o
Windows SharePoint Services Help Search o
Office SharePoint Server Search
Microsoft SharePoint Server 2010/2013:
o
SharePoint Foundation Database o
SharePoint Foundation Help Search (Microsoft SharePoint Server 2010 only)
Data Protector (10.00) Page 287 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution o
SharePoint Server Search o
FAST Search Server 2010 for SharePoint l
The Data Protector Disk Agent component on each Microsoft FAST Search Server 2010 system for SharePoint (Microsoft SharePoint Server 2010, in case of the mixed filesystem + ZDB environment)
Ensure that the Volume Shadow Copy service is started on all these clients.
Installing a medium farm (example)
In Installing a medium farm (example), above
, the Data Protector components that you need to install are colored blue.
Data Protector (10.00) Page 288 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Configuring the integration
For details on how to configure the Data Protector VSS integration, see the HPE Data Protector
Integration Guide for Microsoft Volume Shadow Copy Service.
Configuring user accounts
Create or identify a Windows domain user account that has Windows administrative rights on the
Microsoft SharePoint Server system on which you plan to execute the Data Protector commands. This user must also be granted Microsoft SharePoint Server administrative rights and must be added to the
Data Protector admin user group.
Backup
To back up Microsoft SharePoint Server data, create backup specifications and start backup sessions using the Data Protector PowerShell command
SharePoint_VSS_backup.ps1
.
Prerequisites
l
The Windows Remote Management service (which is used for starting and stopping Windows services remotely, and suspending and resuming FAST for Microsoft SharePoint Server 2010) must be configured on all systems.
To configure and analyze the WinRM service, execute the winrm quickconfig
command.
For more information, see the Windows Remote Management service documentation.
l
In the case of Microsoft SharePoint Server 2010/2013 which uses Microsoft SQL Server 2008/2012 for storing data, and Remote BLOB Storage (RBS) is used with the FILESTREAM provider, ensure that FILESTREAM access level is set to Full access enabled or Transact-SQL access enabled
.
For details of how to configure RBS and FILESTREAM, see the Microsoft SQL Server 2008 documentation.
Limitations
l
The only supported way to start backup sessions is using the Data Protector PowerShell command.
Starting the backup sessions using the Data Protector GUI or CLI is not supported.
l
Microsoft SharePoint Server 2010: With VSS based solution, the FAST Search index files can also be backed up incrementally when using the Data Protector Disk Agent. For all other Microsoft
SharePoint Server data only
Full
backup type is supported.
Recommendations
l
Use the Data Protector PowerShell command to create backup specifications and not the Data
Protector GUI.
Data Protector (10.00) Page 289 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution l
Use the Data Protector GUI to modify backup specifications (for example, to add backup devices).
l
Use the simple mode for the SQL Server databases. In case you want to use the full mode anyway, ensure that you truncate the transaction logs. Otherwise, you may run out of disk space.
l
Whenever you change the farm configuration, perform a new backup.
l
In case you want to back up the Single Sign-On database, do not forget to back up the encryption key as described in: http://technet.microsoft.com/en-us/library/cc262932.aspx#Section32 .
Otherwise, you will not be able to restore the database.
How the command works
When you execute the Data Protector PowerShell command SharePoint_VSS_backup.ps1
, Data
Protector first queries for information about the Microsoft SharePoint Server environment. Then it creates backup specifications.
The newly created backup specifications are named SharePoint_VSS_backup_ClientName and have the same backup device specified for use (the one that you specified at command runtime).
Once the backup specifications are created, the command starts backup sessions (one session for each backup specification).
Microsoft SharePoint Server 2010
In a Microsoft SharePoint Server 2010 environment, the command creates a separate backup specification for each Microsoft SharePoint Server system that has at least one of the following services enabled: l
SharePoint Foundation Database l
SharePoint Foundation Help Search l
SharePoint Server Search 14 l
FAST Search Server 2010 for SharePoint (FAST Search)
For a system with the SharePoint Foundation Database service enabled, the command creates a backup specification that has the SqlServerWriter (Microsoft SQL Server 2008) object selected.
For a system with the SharePoint Foundation Help Search and SharePoint Server Search services enabled, the command creates a backup specification that has the SPSearch4 VSS Writer and
OSearch14 VSS Writer
objects selected.
For a system with the
FAST Search Server 2010
service enabled, the command with the
-hardware option specified creates a VSS backup specification that has the complete FASTSearch home folder
(including bin and lib ) selected.
Microsoft SharePoint Server 2013
In a Microsoft SharePoint Server 2013 environment, the command creates a separate backup specification for each Microsoft SharePoint Server system that has at least one of the following services enabled:
Data Protector (10.00) Page 290 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution l
SharePoint Foundation Database l
SharePoint Server Search 15
For a system with the SharePoint Foundation Database service enabled, the command creates a backup specification that has the SqlServerWriter (Microsoft SQL Server 2008/2012) object selected.
For a system with the SharePoint Server Search services enabled, the command creates a backup specification that has the OSearch15 VSS Writer objects selected.
Considerations
l
The command options enable you to split the process into two parts: first you create the backup specifications and then you start backup sessions. In this way, you can manually modify the newlycreated backup specifications in the Data Protector GUI before the backup is actually started.
l
If Microsoft SQL Server instances are used not only by Microsoft SharePoint Server but also by other database applications, modify the backup specifications so that only the databases that
l
If you have Microsoft SQL Server database mirroring enabled, a failover can occur and so a different
Microsoft SQL Server system becomes active. Since the command creates backup specifications only for the currently active Microsoft SQL Server systems, it is advisable to update (recreate) the backup specifications before the backup is started.
The command syntax
SharePoint_VSS_backup.ps1 –help |-version
SharePoint_VSS_backup.ps1 –createonly
CreateOptions
SharePoint_VSS_backup.ps1 –backuponly
BackupOptions
SharePoint_VSS_backup.ps1 –resumefarm [-preview] | -resumecert
CreateOptions
{-device DevName | -hardware {no_keep|keep|ir} [-device DevName]}
[-overwrite]
[-prefix PrefixName]
[-excludeindex]
BackupOptions
[-outfile
PathToFile]
[-prefix PrefixName]
[-preview]
[-snapshot {diskonly | disktape | tapeonly}]
[-reduce]
[-mode {full | incremental | incremental1 ... | incremental9}]
[-timeout
Timeout]
IMPORTANT:
l
The command must be executed from the Data_Protector_home\bin directory on the front-
Data Protector (10.00) Page 291 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution end Web Server system. Ensure that you are logged in under a user account that is
configured as described in Configuring user accounts, on page 289
and that you open the command prompt with administrative rights.
l
Do not close the PowerShell console while the backup session is in progress. If you close the console during the backup, some actions are not performed: the backup sessions started do finish, but the farm does not resume the original state. To resume the farm, first execute the command with the –resumefarm option and then unquiesce the farm manually using the
Microsoft SharePoint Server Central Administration or stsadm .
Option description
-help
-version
–createonly
–backuponly
-deviceDevName
Displays the SharePoint_VSS_ backup.ps1
command usage.
Displays the SharePoint_VSS_ backup.ps1
version.
If this option is specified, Data
Protector only creates backup specifications. Backup is not started.
If this option is specified, Data
Protector only starts backup sessions using the existing backup specifications. The -device option is not required.
Specifies which Data Protector device to use for backup. You can specify only one device.
IMPORTANT:
If only one device is used to back up a multi-system farm, the corresponding backup sessions cannot run in parallel. This prolongs the time during which the farm is in read-only mode.
Specifically, the farm is in readonly mode from the moment when the backup sessions are started up until all VSS snapshots are created.
To enable backup sessions to run in parallel, select different or additional devices in each backup specification before the backup is
started. See the section Modifying
Data Protector (10.00) Page 292 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
-hardware {no_ keep|keep|ir}
–overwrite
-prefixPrefixName
-outfilePathToFile
backup specifications, on page
.
Specifies that the hardware provider should be used (instead of the software provider with
-device
option specified) and, consequently, ZDB options set.
The default values for ZDB options are as follows: l
Keep the replica for instant recovery: selected if ir is specified.
l
Keep the replica after the backup: selected if ir or keep is specified.
l
Configuration check mode: Strict l
Replica type: Mirror/Clone (Plex) l
Numbers of replica rotated: 3
The default ZDB backup types are as follows (provided a device is also specified): l no_keep : ZDB-to-tape l keep : ZDB-to-disk+tape l ir : ZDB-to-disk+tape
By default, Data Protector does not create backup specifications if they already exist. If this option is specified,
Data Protector overwrites the existing backup specifications with the newlycreated ones. Not applicable if – backuponly is specified.
With this option specified, the backup specifications are created under a different name: SharePoint_VSS_ backup_PrefixName_ClientName .
In case of backup, this option specifies which backup specifications to use: those which name contains
PrefixName.
Non-ASCII characters in PrefixName are not supported.
If this option is specified, backup specification names, errors, sessions outputs, and omnir restore commands are written to the specified file.
Data Protector (10.00) Page 293 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
-preview
-snapshot
{ diskonly
|disktape|tapeonly}
-reduce
-excludeindex
-mode
{ full
| incremental
|incremental1...
|incremental9}
If this option is specified, Data
Protector displays information about the
Microsoft SharePoint Server environment and describes the related actions without actually performing them.
Applicable when starting ZDB backup sessions (that is, sessions that use backup specifications in which a hardware provider is specified for use).
Performs a ZDB-to-disk ( diskonly ),
ZDB-to-tape ( tapeonly ) or ZDB-todisk+tape ( disktape
) session.
Microsoft SharePoint Server 2010: If this option is specified, the command excludes mirrored query components from backup to reduce the backup size.
Microsoft SharePoint Server 2013: If this option is selected, the command selects primary index replicas of each index partition to reduce the backup size.
Applicable only to a Data Protector standard filesystem backup of the
FAST Search index files (Microsoft
SharePoint Server 2010/2013). If this option is specified, Data Protector excludes data_index folder contained in the
FASTSearch
home folder from backup specification. This way, the backup is faster, but the restore is more time consuming. The option enables balancing between a backup size and a time to recovery.
Applicable only to a Data Protector standard filesystem backup of the
FAST Search index files (Microsoft
SharePoint Server 2010/2013). With this option specified, either a Full or
Incremental or leveled incremental backup can be started. By default, the
Full
backup is performed.
When the incremental option is specified and the Full backup does not exist, the option is ignored and the Full
Data Protector (10.00) Page 294 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
-resumecert
-resumefarm
-timeout
Timeout
filesystem backup of the FAST Search index files is started.
Applicable only to Microsoft FAST
Search Server 2010/2013. If this option is specified, the FAST Search certificates for the content and the query connectors are reinstalled.
IMPORTANT:
The SharePoint_VSS_ backup.ps1 –resumecert command must be started on the
Microsoft SharePoint Server system where the SharePoint
Server Search 14 service is enabled.
To be used after restore. This option returns the farm to a working state by resuming all background activities and crawling, unlocking sites, and starting
Microsoft SharePoint Server services.
IMPORTANT:
The command with the resumefarm
option specified uses the WMI (Windows Management
Instrumentation) to remotely start any stopped SharePoint services.
To ensure its proper operation, an exception must be added to the
Windows Default Firewall for
Remote administration , which adds the WMI ports, or for the
WMI directly. For details, see: http://support.microsoft.com/kb/1
54596 .
This option sets the timeout in minutes after which the crawl of the FAST
Search index files is aborted and the farm is resumed. If not specified, the default timeout is 15 minutes.
Starting Windows PowerShell
1. Log in to the Microsoft SharePoint Server system where Windows PowerShell and User
Interface
component are installed, under a user account that is configured as described in
Data Protector (10.00) Page 295 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Configuring user accounts, on page 289
.
2. Open the Windows PowerShell CLI. For example:
Start > Programs > Accessories > Windows PowerShell > Windows PowerShell
3. In case you have Windows User Account Control (UAC) enabled, ensure that you open the CLI with administrative rights. Otherwise, you will not be able to run the Data Protector PowerShell command.
4. Ensure that the Windows PowerShell execution policy is set to RemoteSigned
or
Unrestricted
.
Displaying the Data Protector PowerShell command syntax, below shows how the Windows
PowerShell execution policy is set to
Unrestricted
and how the Data Protector PowerShell command syntax is displayed.
Displaying the Data Protector PowerShell command syntax
Data Protector (10.00) Page 296 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Creating backup specifications (examples)
1. To create backup specifications in which the backup device filelib_writer1 is specified for use, execute:
SharePoint_VSS_backup.ps1 –createonly –device filelib_writer1
2. To create backup specifications with the label weekly in their names and in which the backup
Data Protector (10.00) Page 297 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution device dev1 is specified for use, execute:
SharePoint_VSS_backup.ps1 –createonly –device dev1 –prefix weekly
3. To create ZDB backup specifications in which the backup device dev1 and the hardware provider
(ZDB disk array) are specified for use, and in which the ZDB option Keep the replica for instant recovery is enabled, execute:
SharePoint_VSS_backup.ps1 –createonly -hardware ir –device dev1
4. Applicable only to a Data Protector standard filesystem backup of the FAST Search index files
(Microsoft SharePoint Server 2010).
To create filesystem backup specifications in which the backup device dev1
is specified for use and with the data_index
folder, contained in the
FASTSearch
home folder, excluded from the backup of the FAST Search index files, execute:
SharePoint_VSS_backup.ps1 –createonly –device dev1 -excludeindex
Modifying backup specifications
To modify a backup specification, open the Data Protector GUI. In the Context list, select Backup and, under MS Volume Shadow Copy Writers or under Filesystem (if performing a standard filesystem backup of the FAST Search index files), click the name of the backup specification that you want to modify.
Source page
To modify the Source page of the backup specification (for example, you want to back up individual
Microsoft SharePoint Server databases), consider the following: l
The configuration database and the Central Administration content database must both be backed up during the same time period which starts with the suspend of the Microsoft SharePoint Server
(SharePoint farm) and ends with the resume of the SharePoint farm to ensure data consistency.
l
Microsoft SharePoint Server 2010: The Help Search database and the associated index files must all be backed up during the same time period which starts with the suspend of the Microsoft
SharePoint Server (SharePoint farm) and ends with the resume of the SharePoint farm to ensure data consistency.
l
Microsoft SharePoint Server 2010: The FAST search index files and the FAST Content SSA crawl components must all be backed up during the same time period which starts with the suspend of the Microsoft SharePoint Server (SharePoint farm) and ends with the resume of the SharePoint farm to ensure data consistency.
l
Microsoft SharePoint Server 2010/2013: The SharePoint Service Applications, Search database
(
SSA_Search_DB
), and the associated search index files must all be backed up during the same time period which starts with the suspend of the Microsoft SharePoint Server (SharePoint farm) and ends with the resume of the SharePoint farm to ensure data consistency.
Otherwise, after restore, the Microsoft SharePoint Server data may not be consistent.
Destination page
In the Destination page of the backup specification, you can select different or additional devices and set the device and media options.
Data Protector (10.00) Page 298 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Options page
In the Options page of the backup specification, you can modify backup options. For the standard filesystem backup of the FAST search index files leave the Use Shadow Copy option specified to enable the use of the VSS. To modify ZDB options, click Advanced in the Backup Specifications field and then, in the Backup Options dialog box, click the Advanced backup options tab.
Starting backup sessions (examples)
1. To preview the actions that are performed when a backup session is started, execute:
SharePoint_VSS_backup.ps1 –backuponly –prefix dev –preview
2. To start backup sessions using the existing backup specifications that have no prefix in their names, execute:
SharePoint_VSS_backup.ps1 –backuponly
3. To start backup sessions using the existing backup specifications that have the prefix weekly in their names, execute:
SharePoint_VSS_backup.ps1 –backuponly –prefix weekly
4. To start backup sessions using the existing backup specifications that have no prefix in their names and to save the output of the sessions and the associated restore commands to the file c:\logs\shp.log
, execute:
SharePoint_VSS_backup.ps1 –backuponly –outfile C:\logs\shp.log
5. To start ZDB-to-disk backup sessions using the existing ZDB backup specifications that have no prefix in their names, execute:
SharePoint_VSS_backup.ps1 –backuponly -snapshot diskonly
6. To start incremental filesystem backup sessions of the FAST Search index files (Microsoft
SharePoint Server 2010), execute:
SharePoint_VSS_backup.ps1 -backuponly -mode incremental
Scheduling backup sessions
You can schedule backup sessions using the Windows system scheduler.
1. On the front-end Web server system, create a Windows PowerShell scheduled task. Go to:
Start > Settings > Control Panel > Scheduled Tasks > Add Scheduled Task
2. Open advanced properties for the task.
Data Protector (10.00) Page 299 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Scheduling a backup session using the Windows scheduler
In the Run text box, enter:
For details on
Options
, see The command syntax, on page 291
.
In the Start in text box, enter:
In the Run as text box, enter a Windows domain user account DOMAIN\UserName that is configured as
described in Configuring user accounts, on page 289 .
Restore
To restore Microsoft SharePoint Server data: l
Stop Microsoft SharePoint Server services l
Restore the data.
l
Return the farm to a working state.
For details, see the following sections.
Data Protector (10.00) Page 300 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Before you begin
l
Stop and disable the following services: o
SharePoint Server Search 14 (Microsoft SharePoint Server 2010) o
SharePoint Server Search 15 (Microsoft SharePoint Server 2013)
In addition, stop the following services:
Microsoft Office SharePoint Server 2007:
o
Windows SharePoint Services Administration o
Windows SharePoint Services Search o
Windows SharePoint Services Timer
Microsoft SharePoint Server 2010:
o
SharePoint 2010 Administration o
SharePoint Foundation Search V4 o
SharePoint 2010 Timer o
SharePoint 2010 Tracing o
FAST Search for SharePoint o
FAST Search for SharePoint Monitoring
Microsoft SharePoint Server 2013:
o
SharePoint Administration o
SharePoint Search Host Controller o
SharePoint Timer Service o
SharePoint Tracing Service l
Put the Microsoft SQL Server instance offline if you plan to restore one of the following Microsoft
SQL Server databases: o master o model o msdb o a database for which Microsoft SQL Server mirroring is enabled
Data Protector (10.00) Page 301 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
NOTE:
l
If you use SqlServerWriter , you can restore the model and msdb databases also when the
Microsoft SQL Server instance is online. This is one advantage over MSDE writer. l
Microsoft SQL Server mirroring : If the original and mirror database reside in separate
Microsoft SQL Server instances, put offline both Microsoft SQL Server instances.
Restoring data
You can restore Microsoft SharePoint Server data using the Data Protector GUI or CLI.
Considerations
l
The configuration database and the Central Administration content database must both be restored using backups from the same point in time (the backups performed in the same period in which the
Microsoft SharePoint Server (SharePoint farm) was in suspended mode) to ensure the data consistency. Since the configuration database and the Central Administration content database contain system-specific information, you can restore them only to the original environment or to an environment that has precisely the same configuration, software updates, server names, and number of servers.
l
Microsoft SharePoint Server 2007/2010: The Help Search database and the associated index files must all be restored using backups from the same point in time (the backups performed in the same period in which the Microsoft SharePoint Server (SharePoint farm) was in suspended mode) to ensure data consistency.
l
Microsoft SharePoint Server 2010:
o
Since the FAST configuration database and the FAST Search home folder contain systemspecific information, you can restore them only to the original environment or to an environment that has precisely the same configuration, software updates, server names, and number of servers.
o
The FAST Search index files and the FAST Content SSA crawl components must all be restored using backups from the same point in time (the backups performed in the same period in which the Microsoft SharePoint Server (SharePoint farm) was in suspended mode) to ensure data consistency.
l
Microsoft SharePoint Server 2010/2013: The SharePoint Service Applications, Search database
( SSA_Search_DB ) and the associated index files must all be restored using backups from the same point in time (the backups performed in the same period in which the Microsoft SharePoint Server
(SharePoint farm) was in suspended mode) to ensure data consistency.
l
The following table shows which VSS restore modes are supported for which writers:
VSS supported restore modes and writers
Writers VSS restore modes
Restore to another client Restore files to temporary location
Data Protector (10.00) Page 302 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
MSDE writer
SqlServerWriter
No
OSearch VSS writer
OSearch14 VSS writer/OSearch15 VSS writer
Yes
SPSearch VSS writer
SPSearch4 VSS writer
Yes
Yes (manual attach needed)
No
No
Prerequisites
l
Applicable only to a Data Protector filesystem restore of the FAST Search index files (Microsoft
SharePoint Server 2010). Before restoring the FAST Search index files, the Overwrite option must remain selected to ensure the data consistency. It is selected by default.
Restoring using the Data Protector GUI
1. In the Context List, click Restore.
2. In the Scoping Pane, expand MS Volume Shadow Copy Writers, expand the client which data you want to restore, and then click MS Volume Shadow Copy Writers.
If performing a filesystem restore of the FAST Search index files (Microsoft SharePoint Server
2010), expand Filesystem, expand the client which data you want to restore, and then click the filesystem object.
3. In the Source page, select the data that you want to restore.
4. In the Options page, specify the restore options.
5. In the Devices page, select devices to use for restore.
6. Click Restore, review your selection, and click Finish.
Restoring using the Data Protector CLI
You can restore Microsoft SharePoint Server data using the Data Protector omnir command. For details, see the omnir man page or the HPE Data Protector Command Line Interface Reference.
If you specified the –outfile option when you ran backup sessions, you can find the necessary omnir commands in the specified file. The following is an example of the omnir
command from such a file.
omnir -vss -barhost SHP-APP
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server 2005:SQLWriter)/SHP-APP/master"
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server2005:SQLWriter)/SHP-APP/model"
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server 2005:SQLWriter/SHP-APP/msdb"
Data Protector (10.00) Page 303 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server 2005:SQLWriter)/SHP-APP/
WSS_Content_SSPAdminAccounting"
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server 2005:SQLWriter)/SHP-APP/SSP_Accounting"
-session 2011/09/25-13
-tree "/SqlServerWriter(SQL Server 2005:SQLWriter)/
SHP-APP/SSP_Accounting_Search"
Limitations
The omnir command syntax should not contain more than 8191 characters. If you have so many –tree objects that the syntax exceeds 8191 characters, split the objects and run two separate sessions.
After the restore
After the restore:
1. Enable the service Office SharePoint Server Search
,
SharePoint Server Search 14
, or
SharePoint Server Search 15 .
2. Bring the Microsoft SQL Server instances online (if offline).
3. Return the farm to a working state (that is, resume background activities and crawling, unlock sites, and start the Microsoft SharePoint Server services) by executing:
SharePoint_VSS_backup.ps1 –resumefarm
NOTE:
l
The command uses the WMI (Windows Management Instrumentation) to remotely start any stopped SharePoint services. Ensure its proper operation by adding an exception to the Windows Default Firewall for
Remote administration
, which adds the WMI ports,
or for the WMI directly. For details, see: http://support.microsoft.com/kb/154596 .
l
If the FAST Search certificates for the content and query connectors are out of sync, you can reinstall them by executing:
SharePoint_VSS_backup.ps1 -resumecert
Start the command on the Microsoft SharePoint Server system where the SharePoint
Server Search 14
service is enabled.
Troubleshooting
This section lists general checks and verifications, plus problems you might encounter when using the
Data Protector Microsoft SharePoint Server VSS based solution.
For Microsoft Volume Shadow Copy troubleshooting information, see the troubleshooting chapter in the
HPE Data Protector Integration Guide for Microsoft Volume Shadow Copy Service.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
Data Protector (10.00) Page 304 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Before you begin
l
Ensure that the latest official Data Protector patches are installed. On how to verify this, see the
HPE Data Protector Help index: “patches”.
l
For general Data Protector limitations, as well as recognized issues and workarounds, see the HPE
Data Protector Product Announcements, Software Notes, and References.
l
For an up-to-date list of supported versions, platforms, and other information, see https://softwaresupport.hpe.com/manuals .
Checks and verifications
If your browsing, backup, or restore failed: l
Examine system errors reported in the debug.log
file.
l
Check if you can do a filesystem backup and restore on the problematic client. For information, see the HPE Data Protector Help.
After restore, you cannot connect to the Central
Administration webpage
Problem
After restore, when you try to connect to the Microsoft SharePoint Central Administration webpage, an error similar to the following is displayed in your web browser:
Windows Internet Explorer:
Retrieving the COM class factory for component with CLSID (BDEADEE2–C265–11D0–BCED-
00A0C90AB50F) failed due to the following error 800703fa.
Mozilla Firefox:
An unexpected error has occurred.
Action
1. Restart Microsoft SharePoint Server services on all clients in the farm.
2. Open the Internet Information Services (IIS) Manager and restart all application pools.
3. In case an application pool fails to be restarted with the following error:
Cannot Restore Application Pool. There was an error while performing this operation.
wait for a few seconds and then restart the operation.
4. Delete browsing history in your web browser.
5. Log in to the Central Administration webpage.
Data Protector (10.00) Page 305 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Backup fails with the error Failed to resume Service
Windows SharePoint Services Help Search
Problem
When you start backup sessions, an error similar to the following is displayed:
Service Windows SharePoint Services Help Search on host
MOSS07-INDEX
-> Resuming background activity ...
ERROR: Failed to resume Service Windows SharePoint Services Help Search on host MOSS07-INDEX
Web site URL: http://moss07-web:2001
Root title: as
-> Resuming background activity
Action
Execute:
SharePoint_VSS_backup.ps1–resumefarm
After restore, a quiesce operation fails
Problem
After you have restored the configuration database and executed SharePoint_VSS_backup.ps1– resumefarm , the data in the Microsoft SharePoint Server file system caches on front-end Web Server systems is not consistent with the data in the newly-restored configuration database. When you try to quiesce the farm, the operation fails with the following error:
An unhandled exception occurred in the user interface. Exception
Information: An update conflict has occurred, and you must re-try this action. The object SessionStateService Parent=SPFarm Name=< farm_config_database_name > is being updated by < domain\username
>, in the w3wp process, on machine < servername >. View the tracing
log for more information about the conflict.
Action
Clear the Microsoft Office SharePoint Server file system cache on all server systems in the farm.
For details, see: http://support.microsoft.com/kb/939308 .
Data Protector (10.00) Page 306 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
After restore, you cannot connect to the FAST Search
Server
Problem
After restore, when you try to connect to the Microsoft FAST Search Server 2010 system for
SharePoint, the operation fails.
FAST Query SSA search operations display an error similar to the following:
The search request was unable to connect to the Search Service .
Action
Execute:
SharePoint_VSS_backup.ps1 -resumecert
NOTE:
The VSS based solution copies the FAST Search certificate FASTSearchCert.pfx
from the
FAST Admin Server system to the SharePoint Server system and installs it. Also, the
SharePoint certificate is copied and installed to each FAST Search Server system. For details, see: http://technet.microsoft.com/en-us/library/ff381244.aspx
.
The SharePoint_VSS_backup.ps1 script stops responding and the farm stays in read only mode
Problem
When starting a backup, the SharePoint_VSS_backup.ps1
script stops responding when a crawl of the Microsoft SharePoint Server is being performed. The issue can appear due to external conditions such as a corrupted SSA index, the need to reissue the certificate manually and so on.
As a result, the farm stays in read-only mode.
Action
Normally, the crawl should be aborted automatically after 15 minutes. If this does not happen:
1. Abort the script by pressing Ctrl+C.
2. Manually resume the farm.
You can specify a different timeout after which the crawl is aborted and the farm is resumed by using the -timeout option.
Data Protector (10.00) Page 307 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
SharePoint Search service application not operational after restore
Problem
After restoring the SharePoint Search service application (SSA) and resuming the SharePoint Server
2013 farm, the SSA status reads Paused for:External request instead of Running , indicating that the SSA is not operational.
Action
Perform the following steps:
1. Using the SharePoint Online Server Management Shell, export the SharePoint Search service application (SSA) topology:
$ssa = Get-SpenterpriseSearchServiceApplication -Identity "NameOfSSA"
Export-SPEnterpriseSearchTopology -SearchApplication $ssa -Filename
"TopologyFilename.xml"
2. Record the SSA application pool identify if it exists:
$ssaAppPool = $ssa.ApplicationPool
If it does not exist, create it by executing:
$ssaAppPool = New-SPServiceApplicationPool -name "ApplicationPoolName" -account
"Domain\Username"
3. Delete the SSA by executing:
$ssa = Get-SPEnterpriseSearchServiceApplication -Identity "NameOfSSA"
Remove-SPEnterpriseSearchServiceApplication -Identity $ssa -RemoveData
Remove-SPEnterpriseSearchServiceApplicationProxy -Identity "NameOfSSAProxy"
4. Using the Data Protector Microsoft Volume Shadow Copy Service integration, restore the SSA databases with the Microsoft SQL Server VSS Writer. Ensure the database names start with the name of SSA.
5. Using the SharePoint Online Server Management Shell, restore the SSA itself by executing:
Restore-SPEnterpriseSearchServiceApplication -Name "NameOfSSA" -ApplicationPool
$ssaAppPool -TopologyFile "TopologyFilename.xml" -KeepId
6. Create the SSA proxy by executing:
$ssa = Get-SPEnterpriseSearchServiceApplication -Identity "NameOfSSA"
New-SPEnterpriseSearchServiceApplicationProxy -Name "NameOfSSAProxy" -
SearchApplication $ssa
7. Stop the SharePoint Search Host Controller service on all systems where the SSA indexing components are installed. Execute the command:
Stop-Service SPSearchHostController
8. Using the Data Protector Microsoft Volume Shadow Copy Service integration, restore the SSA index files with the OSearch VSS Writer.
9. Using the SharePoint Online Server Management Shell, start the SharePoint Search Host
Data Protector (10.00) Page 308 of 413
Zero Downtime Backup Integration Guide
Chapter 5: Data Protector Microsoft SharePoint Server Server VSS based solution
Controller service on all systems where the SSA indexing components are installed:
Start-Service SPSearchHostController
10. Resume the SSA by executing:
$ssa = Get-SpenterpriseSearchServiceApplication -Identity "NameOfSSA"
Resume-SPEnterpriseSearchServiceApplication $ssa
Data Protector (10.00) Page 309 of 413
Chapter 6: Data Protector Virtual
Environment ZDB integration for VMware
Introduction
This chapter explains how to configure and use the Data Protector Virtual Environment ZDB integration for
VMware. It describes concepts and methods you need to understand to back up and restore the VMware vSphere virtual environment that is using NetApp or HPE 3PAR storage.
The Data Protector zero downtime backup (ZDB) functionality offers online backup capabilities with minimal degradation of the application system performance. The load on the application system is significantly reduced because backup is nondisruptive and does not require downtime for virtual machines. The tape backup of the copied data is performed using a separate backup system.
Data Protector Virtual Environment ZDB integration for VMware supports environments where ESX(i) Server systems are set up with NetApp and HPE 3PAR storage system, and managed through a vCenter Server
(vCenter environments).
Data Protector supports: l
NetApp storage: ZDB to tape l
3PAR storage: ZDB to disk, ZDB to tape, and ZDB to disk+tape
During the creation of a replica, the application system is in full operation and VMware virtual environment is actively used. The streaming of the data to tape media is subsequently performed on the backup system. The backed up data can be restored using standard Data Protector restore from tape.
Data Protector offers interactive and scheduled ZDB to tape and standard restore methods.
Supported disk arrays and disk array configurations
Supported array
NetApp storage systems
Supported configurations
Systems running Data ONTAP 8.2 for 7-Mode
Systems running Data ONTAP 8.2 for C-Mode
Systems running InForm OS version 3.1.2, 3.1.3, and 3.2.1
HPE 3PAR storage systems
This chapter provides information specific to the Data Protector Virtual Environment ZDB integration for
VMware. For general Data Protector procedures and options, see the HPE Data Protector Help. For details on ZDB terminology, ZDB types, advantages of offline and online backups, see the HPE Data Protector
Concepts Guide. For more information on Data Protector virtual environment integration, see the HPE Data
Protector Integration Guide. For more information on non–HPE Storage Arrays and HPE 3PAR, see the HPE
Data Protector Zero Downtime Backup Administrator's Guide. For other limitations and recommendations, see the HPE Data Protector Product Announcements, Software Notes, and References.
Data Protector (10.00) Page 310 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Recommendations
l
HPE recommends not to install any Data Protector components on VMware vCenter Server system or VMware ESX(i) Server system.
Integration concepts
Supported environments
Data Protector supports environments where ESX and/or ESXi Server systems (ESX(i) Server systems) are managed through a vCenter Server (vCenter environments). Environments with standalone ESX(i) Server systems and mixed environments, in which some of the ESX(i) Server systems are managed through a vCenter Server system and some are standalone, are also supported.
You can even have multiple vCenter Server systems in your environment, each managing its own set of ESX(i) Server systems.
However you set up your virtual environment, for the Data Protector Virtual Environment ZDB integration for VMware it is required that the ESX(i) Server systems are connected to the same storage system with Fibre Channel. Data Protector Virtual Environment ZDB integration for VMware is only supported in virtual environments running on NetApp, and HPE 3PAR storage system.
VMware ZDB integration architecture below shows the architecture of the Data Protector Virtual
Environment ZDB integration for VMware.
Data Protector (10.00) Page 311 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
VMware ZDB integration architecture
Virtual environment:
In this figure, the virtual environment is a vCenter system managing four ESX(i) Server systems. The
ESX(i) servers are connected to the storage system through Fibre Channel SAN. The first three ESX(i) servers are active application systems, and the fourth one is a mount host, a system dedicated to mounting replicas, used for ZDB backup purposes only.
NOTE:
HPE recommends to use a dedicated ESX(i) Server for ZDB backup purposes (mount host).
The NetApp and HPE 3PAR storage system enables snapshot replication of the disk volumes used by virtual machines. The volumes containing the source or original data objects (S) are replicated to an equivalent number of target volumes (T). When the replication process is complete, the data in the target volumes constitutes the snapshot replica. For more information on disk replication, see the HPE
Data Protector Concepts Guide.
The vCenter system is imported in the Data Protector Cell as VMware vCenter client.
Backup system:
The backup system is used to start the backup process and perform a backup of the resolved replicas to tape. It is recommended to use a dedicated physical backup system.
Data Protector (10.00) Page 312 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Data Protector components:
The following Data Protector components should be installed to enable the Data Protector Virtual
Environment ZDB integration for VMware environment: l
ZDB Integration agent ( NetApp Storage Provider or HPE P6000 / HPE 3PAR SMI-S Agent )
NOTE:
The storage providers plug-ins enable ZDB integration within the Data Protector ZDB SMI-S
Agent.
l
Data Protector Virtual Environment Integration (VEAgent) l
Data Protector Disk Agent l
Data Protector General Media Agent
The ZDB Integration agent, the Data Protector
Virtual Environment Integration
component, and the Data Protector Disk Agent must be installed together on at least one Data Protector client in the cell. This client is called the backup system.
The Data Protector General Media Agent must be installed on clients that will transfer data to backup devices. It can be installed on the backup system.
Restoring virtual machines using Data Protector Virtual Environment ZDB integration for VMware is the same as restoring from a "local or network" backup type.
Backup process
VMware ZDB backup flow
See the HPE Data Protector Concepts Guide for a general description of ZDB concepts.
Data Protector (10.00) Page 313 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
See the HPE Data Protector Zero Downtime Backup Administrator's Guide for a general description of the ZDB-to-tape session flow and for the explanation of actions triggered by ZDB options.
This section provides only the information relevant to the Data Protector Virtual Environment ZDB integration for VMware.
Operations on a replica described below are dependent on or triggered by ZDB options. See the HPE
Data Protector Zero Downtime Backup Administrator's Guide for more information on these options.
1. When you start zero downtime backup of your virtual environment, Data Protector starts the
VEAgent and ZDB Agent on the backup system. VEAgent on the application system resolves the selected objects for backup to virtual machines and creates virtual machine snapshots.
2. VEAgent resolves the selected objects for backup to datastore WWNs and ZDB Agent resolves
WWNs to LUNs.
3. ZDB Agent creates replicas of the resolved LUNs.
4. The replicas are presented to the ESX(i) mount host and backup system backs up all resolved virtual machines to tape.
5. When backup is completed, the replicas are unpresented and deleted.
Backup concepts
What is backed up?
Using the Data Protector Virtual Environment ZDB integration for VMware, you can back up the following VMware objects:
VMware vSphere: l
Virtual machines l
Virtual machine disks l
Virtual machine templates
Virtual machines
When you back up a virtual machine, you actually back up virtual machine files of the following types: l
.vmx
l
.vmdk
Virtual machine disks
Data Protector supports backup of individual virtual machine disks when using a vStorage Image backup method. In this case, all virtual machine files are backed up, except for virtual machine disks that are not specified. You can run full, incremental, and differential backups.
From Data Protector 9.05 onwards, the virtual machine disks are backed up in parallel and not serially.
To achieve the virtual machine disk parallelism, the virtual machine disks are considered as objects.
When the object operations (such as object copy and object verification) are specified, the disk objects are not shown. The disks are considered when the virtual machine object is selected for the object operations.
Data Protector (10.00) Page 314 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
NOTE:
After you add a new disk to a virtual machine, make sure you run a full backup session for the updated virtual machine.
Virtual machine templates
You can also back up virtual machine templates when using a vStorage Image backup method. When you create a backup specification, expand the vm folder and select the desired virtual machine templates.
vStorage Image backup method
The vStorage Image backup method provided by the Data Protector Virtual Environment ZDB integration for VMware is based on the VMware vStorage technology. For this method, a single central
backup host is used to back up all virtual machines hosted by ESX(i) Server systems in a Data
Protector cell.
During a ZDB backup, VEAgent performs vStorage Image backup. VEAgent first establishes connections between the backup system and the virtualization host (an ESX(i) Server system). This connection can either be through a vCenter Server system (in the case of a vCenter environment), or direct (in the case of a standalone ESX(i) Server environment). It then requests a snapshot of the virtual machine that is to be backed up, via the vStorage API for Data Protection (VADP). This ensures that the virtual machine is in a consistent state. ZDB Agent then creates a replica of the virtual machine disks. The replica is presented to the ESX(i) mount host, Media Agent client is initialized and the data for backup is streamed to tape.
Snapshot management
The vStorage Image backup method relies on being able to create virtual machine snapshots. A virtual machine snapshot is an operation that puts the virtual machine into a consistent state. All subsequent changes made to the virtual machine disks are recorded to the created snapshot.
NOTE:
The snapshot operation is not supported by all virtual machine disks. For example, snapshots of independent disks are not supported; consequently, this type of virtual disks cannot be backed up using the Data Protector Virtual Environment ZDB integration for VMware. For details, see the VMware documentation.
During a vStorage Image backup, Data Protector creates a snapshot, replicates virtual machine disks and copies the data from the consistent state to Data Protector media. Then Data Protector deletes the replica and the snapshot. Note that snapshots created by Data Protector (DP snapshots) are distinguished from other snapshots by the label
_DP_VEPA_SNAP_
containing the product name, a description, and a timestamp.
NOTE:
If a virtual machine has a user-created snapshot during CBT enabled backup, the user-created snapshot is backed up along with the other VM disk blocks. If the Data Protector detects a usercreated snapshot at the time of restore in a virtual machine, the VM is not restored. To restore such a VM, you need to manually delete all the existing user-created snapshots.
Data Protector (10.00) Page 315 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Virtual machine snapshot tree
Existing virtual machine snapshots reduce the overall performance of a virtual machine. For this reason, Data Protector automatically removes DP snapshots once they are no longer needed.
IMPORTANT:
Do not use the label _DP_VEPA_SNAP_ for snapshots that you create for other purposes, otherwise they will be deleted by Data Protector.
The replicas created by the ZDB Agent are always deleted after ZDB backup is completed.
Backup types
The type of backup to be performed is specified at backup specification level, either in the Scheduler page, or in the Start Backup dialog box for an interactive backup.
Using the vStorage Image backup method, you can perform the following backup types:
Backup types
Backup Types
Full
Description
Backs up the complete virtual machine (disk).
Incremental Backs up the changes made to a virtual machine (disk) since the last full, incremental, or differential backup.
Differential Backs up the changes made to a virtual machine (disk) since the last full backup.
For incremental and differential backup sessions, you must also specify how Data Protector should identify changes at disk block level.
To identify changes at disk block level, Data Protector uses the VMware changed block tracking
functionality. For details, see Changed block tracking, on the next page .
Data Protector (10.00) Page 316 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
NOTE:
The number of snapshots remaining after backup is always 0. Only the Mixed snapshot handling mode is supported.
Mixed snapshot handling mode supports full, differential, and incremental backups in all possible backup chain forms.
For detailed description of snapshot handling modes, see HPE Data Protector Integration Guide.
When performing snapshot operations on a backed up virtual machine, you must be careful not to break your backup chains.
IMPORTANT:
A VMware object’s snapshot which was not created with Data Protector cannot be used to set up a Data Protector backup chain (restore chain) for that object.
A backup chain gets broken if you perform any of the following operations: l
Delete a snapshot l
Revert to a snapshot l
Create a snapshot without involving Data Protector l
Change snapshot handling mode l
Add a new virtual machine disk or rename an existing one l
Restore a virtual machine l
Enable changed block tracking
After completing any of the above operations, you must first run a full backup to start a new backup chain. If you run a session for incremental or differential backup instead, the Data
Protector switches the VEAgentdisk objects to have the effective backup type as full, whereas for VEAgent object, the backup type still remains as incremental or differential. This may create a backup chain with multiple sessions during restore; which impacts performance. So, it is recommended to execute a full backup.
Changed block tracking
Changed block tracking (CBT) is a feature of later versions of VMware that can be used to improve the efficiency and speed of your backups.
For CBT, change IDs are used. A change ID is an identifier for the state of a virtual disk at a specific point in time. It is saved by the virtual disk logic whenever a snapshot is taken of the disk.
The main advantage of using changed block tracking is most noticeable on incremental or differential backups, because: l
It is not necessary to keep virtual machine snapshots on the system until the next backup, greatly reducing the system overhead.
l
The changes to be backed up are calculated more easily, by obtaining change information from the kernel, rather than calculating it from snapshots.
During full backups, only active blocks on the disk are backed up, and unallocated blocks are ignored.
This makes the backups space-efficient and faster.
Data Protector (10.00) Page 317 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
When changed block tracking is enabled, a virtual machine's performance is slightly impacted, but this is small relative to what you gain. You can enable it using the Data Protector GUI, if required.
NOTE:
The CBT functionality is supported for vStorage Image + OpenStack backup method.
When changed block tracking is used, Data Protector snapshots are still used to put a virtual machine in a consistent state. However, when the backup completes, they are deleted. Only changed block tracking log files change IDs are kept.
NOTE:
When working with changed block tracking, keep a note of the following points: l
Ensure that the prerequisites from VMware are met while using CBT backup. For more information, go to https://kb.vmware.com/kb/1020128 l
Not all types of virtual disk support changed block tracking. In case of an unsupported disk, a virtual machine backup fails.
l
When changed block tracking is first enabled, the next backup for a virtual machine will always be a full backup to provide a reference point for the tracking. In other words, a new backup chain is started.
l
A CBT backup chain (Full, Differential, Incr,...) gets broken when you perform a restore session. When the restore session completes, run a full backup again to start a new backup chain, otherwise subsequent incremental and differential backup sessions will fail.
Backup flow
1. Data Protector triggers a snapshot.
2. Replicas of source volumes are created.
3. A change ID is recorded for the current backup.
If this is the first snapshot taken after changed block tracking was enabled, all active blocks are identified and change ID 0 is recorded.
In case of a full backup, this change ID becomes the start reference point for a new backup chain.
4. This step depends on the backup type selected: l
Full: The blocks changed since change ID 0 are identified.
l
Incremental: The blocks changed since the change ID for the previous backup (full, incremental, or differential) are identified.
l
Differential: The blocks changed since the change ID for the previous full backup are identified.
5. The identified blocks are backed up.
6. The replica and the snapshotare deleted.
Example of a backup chain with changed block tracking
Snapshot Change ID
1st after CBT enabled ID 0
Blocks identified
All active blocks
Blocks backed up
—
Data Protector (10.00) Page 318 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Snapshot
Full backup
Change ID
ID n
Incremental backup ID n+m
Incremental backup ID n+p
Differential backup ID n+q
Full backup ID r
Blocks identified
Changed since ID 0
Changed since ID n
Changed since ID n+m
Changed since ID n
Changed since ID 0
Blocks backed up
All active blocks from
ID 0 + changed blocks since ID 0
Changed block since
ID n
Changed blocks since
ID n+m
Changed blocks since
ID n
All active blocks from
ID 0 + Changed blocks since ID 0
Data Protector (10.00) Page 319 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Non-Changed Block Tracking (Non-CBT) backup
Non Changed Block Tracking (Non-CBT) backup is a feature which does not depend on block level changes for backup. l
In this feature, all the blocks of the virtual machine disks are backed up.
l
This feature does not use the VMware CBT functionality to identify the modified blocks for backup.
l
The size of the backed up image increases, as all the blocks of disk are backed up.
Allow fallback to non-CBT backups option is enabled when change block tracking backup fails.
The non-CBT backup can be used in the following scenarios: l
When the hardware version of virtual machine is lesser than 7.
l
When backing up virtual machines without the older version of operating system installed (example,
Windows 2003).
l
When snapshots are available on virtual machine and CBT is not enabled.
Limitations
Data Protector (10.00) Page 320 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
Only full backups are available for non-CBT backups. l
Incremental or differential backups are not supported.
Quiescence
If quiescence is selected, the snapshot process quiesces all the system writers, and the registered application writers. In Windows guest operating systems, the VSS framework freezes or quiesces the applications running in a virtual machine before a snapshot is created. Data Protector performs application-consistent quiescence every time the Use Quiescence option is selected for the backup.
The illustration below shows the options available for the Quiescence functionality to quiesce applications with VSS writers.
Select the Use quiescence snapshots check box. You can select the level of error to be reported. The following levels of errors can be selected:
Data Protector (10.00) Page 321 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
Fatal: If the quiescence snapshot fails, the session will fail . l
Warning: If the quiescence snapshot fails, a warning message is displayed and the backup process is continued.
For CLI options, see the CLI Reference Guide. Navigate to Section 1M: Administrative commands >
vepa_util.exe(1M).
NOTE: When you enable the Quiescence option, backup of virtual machines are application consistent for MS SQL, MS Sharepoint, MS Exchange and Oracle. The consistency works best when the applications are configured as recommended by their vendors. Data Protector recommends installing a respective integration agent within a virtual machine to protect these applications.
Prerequisites
l
For VEPA-based quiescence backups of the Share Point server, you must register the VSS Writer before starting the backup of virtual machines. For more information, see the Integration guide for
Microsoft Volume Shadow Copy Service, Microsoft SharePoint Services writer specifics section. l
For VEPA-based quiescence backups of the Oracle 11g Release 2 database, the Oracle VSS writer service should be in the active/running state, before starting the backup of virtual machines. This can be done by executing the command oravssw /q /start . For other applications such as SQL and SharePoint, the VSS writers are registered and active by default.
Limitations
l
The Quiescence feature can slow down the speed of backup sessions considerably.
l
The Microsoft SQL database on the Availability Group cluster application in the Windows OS cluster is not supported for VEPA quiescence.
l
The cluster applications using SCSI controllers for disks in shared mode, are not supported for
VEPA backups.
l vRDM disks can be used only for full backups.
l
Quiescence backups for virtual machines in the Power Off state are not valid.
Considerations for Quiescence Operations
VMware considerations
l
Do not disable the UUID attribute for virtual machines. l
The virtual machine must use only SCSI disks. Application-consistent quiescing is not supported for virtual machines with IDE disks. There must be as many free SCSI slots in the virtual machine as the number of disks.
l
Physical RDM's don't support snapshots, therefore they cannot be used for quiescence.
l
Quiescence is not supported for backups of virtual machines on the Microsoft cluster. For more information, see Backup on virtual machines configured with bus-sharing .
Virtual machine considerations
Data Protector (10.00) Page 322 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
The virtual machine must not use dynamic disks.
l
Ensure that the latest version of VMware Tools are installed in the virtual machine. For more information, see Verifying a VMware Tools build version .
l
VSS components must be explicitly specified during the VMware Tools upgrade process. VSS will not be installed in a non-interactive mode. For more information, see Installing the Volume Shadow
Copy Service with VMware Tools .
l
Ensure that you are using Windows Server 2008 or higher. Previous versions of Windows, such as
Windows XP and Windows 2000, do not include VSS and rely on the SYNC driver.
l
The Distributed Transaction Coordinator service must be running during the installation of VMware
Tools. Otherwise, VSS fails to quiesce Windows 2008 R2/ Windows 2012.
l
Ensure that all the appropriate VSS application services are running, and the startup types are listed correctly.
For more information on VSS quiesce related issues, see Troubleshooting Volume Shadow Copy
(VSS) quiesce related issues .
Disk space requirements
A virtual machine backup requires sufficient disk space for snapshots and replicas on a storage system where the virtual machine disks reside.
Two replica provision types are available for managing storage resources: l thin provisioned
With thin provisioning you can conserve storage resources. The disk space allocation is dynamically adapted to accommodate the demand for storage. It allows free space sharing between LUNs and enables LUNs to consume only the space they actually use.
The required free space is based on the delta file created by a Data Protector snapshot. Only changes made since the last backup session are written into the replica l fully allocated
With this provision type space-reserved LUNs and snapshot copies have pre-allocated space that can be continually overwritten. This guaranteed space is not available to any other LUNs or snapshot copies within the volume.
The required free space is calculated based on the size of virtual machine disks just before the snapshot is created. The overall disk space required is double the space required for snapshots. A replica requires as much space as the source LUN. If the target is not space-reserved, ensure that the volume has enough free space to accommodate the replica.
Free space required option
You can use the Data ProtectorFree space required % option to make sure that a virtual machine is backed up only if there is enough free space.
The required free space is calculated based on the size of virtual machine disks just before the snapshot is created. Data Protector checks all datastores where the virtual machine disks reside. If one of the datastores does not meet the specified percentage of free space, no snapshot is created and the backup of the virtual machine fails with an error.
Data Protector (10.00) Page 323 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
When backing up more than one virtual machine, the check is applied to each virtual machine separately. The virtual machines which pass the check are backed up and the ones which do not pass it are not.
If you specify 0%, the check is omitted.
Examples
The following examples illustrate how the Free space required % option works:
1. Backup of a single virtual machine "test1" with the disk "disk1" residing on the datastore
"datastore1":
If you specify 30% in the Free space required % option and the datastore has a size of 100 GB, the backup succeeds if there is at least 30 GB of free space on the datastore.
2. Backup of a single virtual machine "test1" with two disks, "disk1" and "disk2" residing on two datastores, "datastore1" and "datastore2":
If you specify that 30% of free space is required, the backup succeeds if there is at least 30 GB of free space on each datastore.
3. Backup of two virtual machines, the virtual machine "test1" with the disk "disk1" on the datastore
"datastore1" and the virtual machine "test2" with the disk "disk2" on the datastore "datastore2":
If you require 30% of free space, the backup of both virtual machines succeeds if there is at least
30 GB of free space available on each datastore. If for example, the datastore "datastore1" has less than 30% of free space and the datastore "datastore2" has at least 30% of free space, the backup of the virtual machine "test1" fails and the backup of the virtual machine "test2" succeeds.
If both datastores have less than 30% of free space, the backup of both virtual machines fails.
Disk space requirements
Backup methods
vStorage
Image
Required disk space on datastores
Explanation
The sum of the sizes of all the virtual machine disks, plus: l
The size of any quiescence zip files, if quiescence is specified.
When a virtual machine snapshot is taken, changes made to the virtual machine disks are recorded to separate files (one delta file is created for each virtual machine disk). A delta file can grow up to the original virtual disk size.
Backup disk buffer
You can specify a disk buffer for your backup using the omnirc option OB2_VEAGENT_BACKUP_DISK_
BUFFER_SIZE .
The SAN and the HotAdd backups support disk buffer sizes from 1 MB to 256 MB. By default, their disk buffer size is 8 MB. However, network backups, such as NBD and NBD (SSL), are always performed using the default disk buffer size of 1 MB.
NOTE:
l
If there is not enough memory available for the specified disk buffer size, a fallback to 1 MB disk buffer size will happen to keep the backup running, and a warning message will be
Data Protector (10.00) Page 324 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware displayed.
l
Using bigger disk buffer sizes improves the backup performance, but it also increases the memory consumption. At the certain level the backup performance does not improve anymore due to the limits of your backup host.
Backup parallelism
By default, virtual machines are backed up in parallel. In rare cases this may lead to problems. For example, backup sessions may end unexpectedly. In such cases, you may set the Data Protector
OB2_VEAGENT_THREADED_BACKUP omnirc option on the backup host to 0 to disable parallel backups.
NOTE:
Virtual machine disks are backed up sequentially in both cases.
For details on how to use Data Protector omnirc options, see the HPE Data Protector Help index:
“omnirc options”.
Backup considerations
l
Change Block Tracking (CBT) and Mixed snapshot handling mode
CBT backup method and mixed snapshot handling mode are supported. l
Concurrent backup sessions
Backup sessions that use the same devices cannot run in parallel.
You cannot back up virtual machines in parallel with an ESX(i) Server system or a datacenter where the virtual machines reside.
Backup sessions that backup virtual machines in the same data store cannot be run in parallel either in the same cell manager or different cell manager.
l
Transportation modes
.
It is recommended to use CBT for making incremental/differential backups because it is faster and uses less space on the backup device.
The transportation mode can either be SAN or NBD, and this can be selected from the Data
Protector GUI (This determines how the array will be accessed). The transportation mode can also be configured in the virtual machine options. The transportation mode configured in the virtual machine options is executed on priority, and it follows the order
SAN:HOTADD:NBDSSL:NBD:FILESYSTEM
.
For example, o
If you have selected NBD in the virtual machine options from the Data Protector GUI o
If SAN is not available o
If the HOTADD solution is not possible for the zero downtime backup
Data Protector (10.00) Page 325 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Then the backups will go through NBDSSL. However, if you want the backups to go through NBD transport mode, then you must configure the NBD transport mode in the virtual machine options.
A template backup in the SAN transport mode is not supported, and will fall back to the NDB transport mode.
l
Replica provision type
To manage storage resources you can select thin provisioned or fully allocated replica provision type to ensure enough free disk space on a storage system.
l
Presentation of LUN
For zero downtime backups of virtual machines from the 3PAR replica, ensure that the LUN that is used to create the source data store (where the virtual machine to be backed-up resides) is not presented to the system that is configured as the mount proxy host.
l
Backup to StoreOnce Catalyst device
From 9.07 onwards, all VEPA backups to StoreOnce Catalyst device are performed with "Single
Object Per Media Store" mode. Even if this option is not selected on the StoreOnce Catalyst device, this mode will be enforced.
Any value that you enter in the "Store Media Size Threshold (GB)" field will be ignored to enable
Cached GRE or Power On and Live Migrate from the backups done to StoreOnce Catalyst device.
l
Data Protector licenses
The following licenses are not required for performing zero downtime backups of virtual machines from the 3PAR replica: o
HPE Data Protector instant recovery extension for UNIX - 1 TB o
HPE Data Protector instant recovery extension for Linux - 1 TB o
HPE Data Protector instant recovery extension for Windows - 1 TB
Restore concepts
You can restore VMware objects backed up with vStorage Image backup method in different ways.
Restore of VMware objects backed up with vStorage Image method
Virtual machines, virtual machine disks, and virtual machines templates backed up with the vStorage
Image method can be restored: l
To a datacenter l
To a directory on a backup host
Restore to a datacenter
By default, virtual machines are restored to the original datacenter and the original datastore, but you can select a different datacenter if you want.
Data Protector (10.00) Page 326 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
By default, Data Protector deletes a virtual machine (if it exists) before it is restored, even if it resides in a different datacenter from the datacenter you restore to.
NOTE:
If you select an ESX(i) Server client in the Restore client option (destination client) in the restore wizard, the migrated virtual machine will not be deleted, as the ESX(i) Server client is not able to detect virtual machines that are located on different ESX(i) Server clients (only a vCenter client can do that). Consequently, you end up with two virtual machines having the same UUID.
Alternatively, you can choose to restore virtual machines only if they do not exist, leaving existing virtual machines intact.
For the restore, you can also specify the following: l
Whether the restored virtual machines should be registered in the datacenter l
Whether the restored virtual machine snapshots should be consolidated when the restore completes l
Whether the restored virtual machines should be powered on
The restore options provided are, by default, set to restore virtual machines to the original datacenter.
You can restore virtual machines and virtual machine disks, from a replica to a data center. Note that restore sessions from replica to a directory is not supported. In Disk + Tape backups, if the replica is rotated or removed by the administrator, the restore happens from the tape.
Restore of individual virtual machine disks
To be able to restore individual virtual machine disks to a datacenter, the original virtual machine must still exist. Otherwise, the restore fails.
Here is the progress of a restore session:
1. The virtual machine is powered off.
2. If the disks to be restored still exist, they are removed.
3. The disks are restored from the backup.
NOTE:
After restore, virtual disks that are part of a dynamic disk set or virtual disks from different points in time may require additional user action (for example, mounting, resignaturing, or recovery) in the guest operating system and/or applications running inside.
Restore to a directory
When restoring to a directory (restore outside a datacenter) all the files of virtual machines are restored to a directory of your choice (for example,
C:\tmp
) on a backup host.
In the directory you specified, subdirectories are created with names corresponding to those of the datastores where the virtual machines (their virtual disks) resided at the time of backup. The files related to the virtual disks are restored to the respective subdirectories.
After such a restore, the virtual machines are not functional. You need to manually move the restored virtual machine images to an ESX(i) Server system, using the VMware Converter as described in
Recovering virtual machines after restore to a directory, on page 380 .
Data Protector (10.00) Page 327 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Restore of Nova Instances and Shadow VMs backed up with vStorage Image + Openstack method
Progress of a restore session:
1. IDB is queried to obtain the list of Shadow VMs attached to the Nova Instance.
2. vCenter is queried to create the map of Nova Instance and Shadow VMs.
3. Related Shadow VMs are added into the restore object list.
4. vCenter is queried to create the map (
) and validated if the Shadow VMs are attached to the original instance.
5. Queried the restore version from IDB to obtain the restore chain.
6. Shadow VM files are removed from the vCenter.
7. Nova Instance files are removed from the vCenter.
8. Restored the Nova Instance and disk files in the same folder structure.
9. Restored the Shadow VM configuration files.
10. Registered the virtual machine and restored the network.
NOTE:
If the selected Nova Instance is available in the vCenter during the restore, the Shadow VMs are detached before deleting the Nova Instance.
After the Data Protector restore, to bring back the OpenStack instance in the OpenStack Horizon dashboard to the correct state, perform the following steps:
1. Restart the Nova Compute service in the Nova Proxy node by executing the following command:
Service nova-compute restart
2. Refresh the Horizon Dashboard to check if the Nova Instance is available or not. If the Nova
Instance is not available, connect to the OpenStack Management node and execute the following commands: l
Nova list : Lists the Nova Instances.
l
Nova reset-state –active “instance-uuid” : Resets the state of the Nova Instance to active.
l
Nova reboot “instance-uuid” : Reboots the Nova Instance.
Once the reboot is done, refresh the Horizon Dashboard and check for the availability of the Nova
Instance.
Restore chain
When you restore a virtual machine from a backup created in an incremental or differential session,
Data Protector automatically restores the complete backup chain, starting with the last full backup, which is then followed by the differential and all subsequent incremental backups (if they exist) up to the selected session.
Data Protector (10.00) Page 328 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Power On and Live Migrate
Power On
Virtual machines can be powered on instantly within seconds from the Data Protector backup image that resides on the 3PAR replica (local or remote copy), Smart Cache, and StoreOnce Catalyst devices. Previously, virtual machines had to be powered on only after the complete data migration to the production data center. Use this feature if you want to verify the sanity of the backup. Note that the changes done to the virtual machine once it is powered on, will be available until you perform the clean up operation.
When you power on a virtual machine, the backup image is presented to the destination ESX server. A new virtual machine is created, whose data disks point to the Data Protector backup image. The other files reside on the destination data center.
Live Migrate
This option will power on the virtual machine from the backup image, and will simultaneously start the data migration to the destination datastore. During this process, the virtual machine will continue to be accessible. Since the data movement is a back end operation, it will have minimum impact on the usage and accessibility of the powered on virtual machine. Any modifications done to the virtual machine data will be consolidated, and the migrated virtual machine will have all the modified content on top of the restored image from the backup.
Once the data migration is complete, the virtual machine functions from the destination datastore, and has no dependency on the backup image. Also, the backup image presentation is removed.
StoreOnce Catalyst device only
l
Power On and Live Migrate operation from full, incremental, and differential backups are supported.
l
Power On and Live Migrate operation from object copy is supported for 9.05 and 9.06 backups. The object copy should be performed on per session basis to ensure data consistency.
NOTE:
If you want to migrate your Data Protector 9.05 or 9.06 version backups to StoreOnce
Catalyst to use the feature of Power On and Live Migrate, it is recommended to perform object operations at individual session. If you choose multiple sessions at once, data consistency will not be there.
l
If the virtual machine is powered on from the StoreOnce Catalyst device, the virtual machine must be cleaned up before performing the Live Migrate operation.
Cleanup/Power Off
Data Protector stores the list of all the powered on virtual machines. It stores the detailed information about all powered on virtual machines as an XML file in the Cell Server. The cleanup and power off actions are listed below: l
If the virtual machine is already powered on for more than 24 hours, it will be powered off, and the related storage will be cleaned up. l
If the virtual machine is powered on from the replica, the data store will be un-mounted and the replica created during the power on process will be removed from the array. l
If the virtual machine is powered on from the Smart Cache or StoreOnce Catalyst device, during the clean up process, the data store will be removed, and the NFS share will be deleted.
Data Protector (10.00) Page 329 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Note that the above actions are applicable to virtual machines that are powered on using the Power On feature.
The virtual machines that are powered on for more than 24 hours will undergo a cleanup operation during the next daily maintenance job. All the virtual machines that are cleaned up in daily maintenance job are logged in the poweronvms_cleanup.log
file.
For more information on the procedure of Power On and Live Migrate, see Restoring using the Data
.
Note that Power On and Live Migrate is not supported with incremental and differential backups on a
Smart Cache device.
Restore considerations
l
Concurrent sessions
Restore sessions that use the same devices cannot run concurrently.
l
Failed restore sessions
Sometimes, when a virtual machine restore fails, Data Protector creates extra files on the datastore which you need to clean up manually when the session completes. Otherwise, corrupt virtual machine backups may be created in subsequent sessions and restore from a such a backup also
fails. For details, see Cleaning up a datastore after a failed restore, on page 387 .
When restoring a virtual machine to a non-original datastore whose block size is not compatible with the virtual machine disks' sizes (that is, a .vmdk
file size is not a multiple of the datastore block size), the restore fails.
l
Virtual machines in vApp
When you restore a virtual machine that resided in a vApp container at the time of backup, the virtual machine is not restored back to the vApp container, but to the ESX(i) Server root level. If the virtual machine in the vApp container still exists, it is deleted or the restore is skipped, depending on what you select in the Existing virtual machine handling option.
l
Partial restore from a vStorage Image backup
When performing a partial restore from a vStorage Image backup (for example, when restoring only some out of many backed up VM disks), the default option Delete after restore is ignored and the
Delete before restore option is used instead.
l
Transportation modes
The following recommendations for specific virtual machine transportation modes apply: o
SAN transportation mode: To use the SAN transportation mode for restore: n
Select a backup host for a restore session.
n
Ensure that the storage volumes that are presented to both the backup host and ESX(i)
Server systems are not read-only. For details on how to check storage volume properties, see
A restore session using SAN transportation mode fails.
n
Ensure that the storage volume size is a multiple of the underlying VMFS block size.
Otherwise, the Write operation to the remainder fails. For example, if the storage volume size is 16.3 MB and the block size 1 MB, writing to the remaining 0.3 MB fails. For details, see the
VMware Knowledge Base at:
Data Protector (10.00) Page 330 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware http://kb.vmware.com/selfservice/microsites/searchEntry.do
.
Search for “Best practices when using SAN transport for backup and restore”.
o
It is recommended to use CBT for making incremental/differential backups because it is faster and uses less space on the backup device.
o
Hotadd transportation mode: Hotadd transportation mode is available for restore, however
VMware does not support multiple disks. Therefore, in a HotAdd environment use the omnirc option OB2_VEAGENT_RESTORE_TRANSPORT_METHOD to set the restore transport mode to NBD.
Power On considerations
l
Install the following NFS packages on the backup host : o
For Smart Cache backups: NFS version 3 or later o
For StoreOnce Catalyst backups: NFS version 4 or later l
Data Protector uses the Windows PowerShell script nfsServiceCheck.ps1
to initiate the NFS service, which is required for the Power On process. Execution of this script requires the execution policy to be set to RemoteSigned. If you need the policies to be Restricted, set the omnirc variable
OB2_NO_NFSSERVICE_CHECK
to 1. In case the script fails, you need to execute it manually. Some of the possible reasons for the failure of the NFS service installation may be: o
NFS port is used by another application.
o
Powershell may require a reboot.
o
NFS module not present in the Powershell repository.
Use the following command to execute the NFS service manually: o
Windows Command Line: powershell.exe NFSServiceCheck.ps1
o
PowerShell Command Line: NFSServiceCheck.ps1
l
The backups of non-persistent virtual machines that are created as part of the Power On feature will be skipped. The following message is displayed when you try to perform a backup operation: Non
Persistent Powered on Virtual Machine found, Skipping Backup. l
Network will be disabled in the virtual machine that is powered on from the Smart Cache or
StoreOnce Catalyst devices. l
To manually clean up all the VMs that are powered on in your Cell Manager, perform the following:
1. Set the omnirc variable FORCE_PURGE_POWERON_VMS to 1 on all the backup hosts.
2. Run the following command in the Cell Manager:
/opt/omni/sbin/omnidbutil -purge_expired_poweron_vms
NOTE:
All powered on vm's will be powered off and removed.
Data Protector (10.00) Page 331 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
HPE 3PAR storage systems only
l
You must install the vmfs-tools-0.2.5
in the Linux mount proxy host. . l
You must ensure that multipath services are running in the Linux mount proxy host.
l
You must install the sg3_utils rpm package in the Linux mount proxy host.
l
To perform GRE operations from the 3PAR replica, the GRE mount proxy host and the source ESX server must be present in the same 3PAR zone.
StoreOnce Recovery Manager Central Integration
StoreOnce Recovery Manager Central (RMC) software integrates HPE 3PAR StoreServ primary storage with HPE StoreOnce Backup systems. RMC integrates 3PAR StoreServ primary storage and
StoreOnce Backup for converged data protection that delivers assured recovery of applicationconsistent recovery points with flexible recovery options.
The Express Protect feature offers a second-tier of data protection by facilitating direct backup from
3PAR StoreServ to StoreOnce device, independent of backup software. Backups to StoreOnce are self-contained volumes, deduplicated to save space, and can be used to recover back to the original or a different 3PAR StoreServ array, even if the original base volume is lost. The Express Protect feature facilitates direct backup from the primary storage to the backup storage completely removing the application server from the data path.
HPE StoreOnce RMC for VMware enables you to protect VMware Virtual Machine Disks (VMDKs) and data stores using application-consistent snapshots for rapid online recovery.
With Data Protector, you can backup virtual machines snapshots created by RMC to Data Protector supported secondary storage devices. You can then perform a restore operation to the required destination. Note that Granular Recovery Extension (GRE) operations are possible only for
Snapshot+Tape backups.
Data Protector supports the following types of backup: l
Snapshot backups: With Snapshot backups, you can create snapshots of the original volumes. l
Snapshots+Tape: With Snapshot+Tape backups, you can create snapshots, and backup the data to Data Protector supported secondary storage devices. l
Express Protect backups: With Express Protect backups, you can backup snapshots from HPE
3PAR StoreServ to HPE StoreOnce.
The following table lists the supported backup types, backups that support Granular Recovery
Extension, Power On, and Live Migrate operations.
Supported backups
Snapshot
Snapshot+Tape
Backup type
l
Full l
Full
GRE
Not supported
If the tape device is
Smart Cache or
StoreOnce Catalyst, cached GRE is supported. For any
Power On and Live
Migrate
Not supported
Supported, if the tape device is Smart Cache or StoreOnce Catalyst.
Data Protector (10.00) Page 332 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Snapshot+Tape l
Incremental l
Differential other type of tape devices, non-cached
GRE is supported.
If the tape device is
StoreOnce Catalyst, cached GRE is supported. For any other type of tape devices, non-cached
GRE is supported.
Supported, if the tape device is StoreOnce
Catalyst.
Express Protect l
Full l
Incremental
RMC integration considerations
Not supported Not supported l
Single datastores consisting of multiple LUNs are not supported for RMC backups. l
The RMC server details that are provided when you create the backup specification cannot be modified once the specification is saved. l
For RMC integration backups, if automation scripts are used for creating barlists, ensure that the recovery set name, specified for barlist creation, is unique. l
For RMC Express Protect backups, Data Protector maintains 2 snapshots internally. This is required for executing incremental backups. When the snapshot count exceeds 2, Data Protector rotates the older snapshot. l
RMC Express Protect backups are supported only from Inform OS version 3.2.1 MU1 onwards. l
The RMC schedules and backup reports are supported only in Windows and Linux Cell Manager platforms.
l
For RMC restore sessions from snapshot, array credentials must be added using the omnidbzdb command.
l
For RMC restore sessions, only Data Protector tape restores can be restored to a directory.
l
For RMC Snapshot and Express Protect backups, the session information is not displayed in the
GRE GUI.
Prerequisites
You should be familiar with the StoreOnce RMC concepts and procedures. For more information on
RMC, see the HPE StoreOnce Recovery Manager Central User Guide and HPE StoreOnce Recovery
Manager Central for VMware User Guide.
Limitations
Express protect restore: If a restore session is aborted when express protect is restoring data from
StoreOnce to a new snapshot, the Data Protector session is aborted. However, restore from
StoreOnce to a snapshot continues in the background. You need to wait until the RMC restore operation is complete, before another Data Protector restore operation from the same session is triggered.
Data Protector (10.00) Page 333 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
RMC Integration Process
1. Adding the RMC server details: Add the RMC server using the Command Line Interface.
Execute the command omnidb -addhost -servername <clientname> -user <username> psswd <password> .
2. Backup: Perform the backup operation by selecting the Backup Type as StoreOnce RMC in the
Data Protector GUI.
3. Restore: Perform the restore operation or the GRE operation using the GUI.
Configuring the integration
Configure the integration as follows: l
Import VMware clients into the Data Protector Cell.
l
Configure virtual machines you want to back up.
Recommendations
l
HPE recommends not to use the percent sign in names of virtual machines that will be backed up, restored and recovered with Data Protector. If a virtual machine name contains % , the name is displayed incorrectly in the Data Protector GUI and in the Data Protector session messages.
Prerequisites
l
Make sure that you have a correctly installed and configured VMware vSphere environment.
For supported versions, platforms, devices, and other information, see the latest support matrices at https://softwaresupport.hpe.com/ .
l
Ensure you have Disable Methods and Enable Methods Global privileges granted to the user account that is used to connect to the vCenter Server.
l
Make sure you have the necessary VMware vSphere privileges granted to the user account that is
.
l
Make sure that you have correctly installed Data Protector. For information on installing Data
Protector in various architectures, see the HPE Data Protector Installation Guide.
Data Protector Virtual Environment ZDB integration for VMware must be correctly installed and configured. Make sure that you have at least one client with the ZDB Agent (Storage Provider component for the non–HPE Storage Array or the HPE 3PAR Storage Provider component) and
Virtual Environment Integration component installed (backup system) in your environment.
For configuration, see the HPE Data Protector Zero Downtime Backup Administrator's Guide.
IMPORTANT:
The client you intend to use as the backup system should not have the VMware
Consolidated Backup (VCB) software installed.
Data Protector (10.00) Page 334 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
If you intend to restore virtual machine files to a directory on a backup host, also install the Disk
Agent
component on the backup system. Otherwise, you will not be able to use the Browse button to specify the target directory (however, you will still be able to type the directory by yourself).
Before you begin
l
Configure devices and media for use with Data Protector.
Importing and configuring VMware clients
With the Data Protector Virtual Environment ZDB integration for VMware, it is not necessary to install any Data Protector components on the VMware clients (vCenter Server systems, ESX(i) Server systems). To make them Data Protector clients, the VMware clients must be properly imported into the
Data Protector cell and configured.
Procedure
To import a client into a Data Protector cell:
1. In the Context List, click Clients.
2. In the Scoping Pane, expand Data Protector Cell, right-click Clients, and select Import Client.
3. In the Import client page, enter the client name in the Name option, select the appropriate client type (VMware ESX(i), VMware vCenter) from the Type drop-down list and click Next.
Importing a VMware vCenter Server client (Name and Type)
4. If you select Standard security, you need to manually specify which login credentials Data
Protector should use to connect to the VMware client:
Port : Specify the port that VMware vSphere is using. By default, VMware uses the port 443. For
VMware ESX(i) Server that you will use as a mount host, specify the port 443.
Username and Password: Specify an operating system user account that has the following
VMware vSphere privileges on root vCenter level:
Data Protector (10.00) Page 335 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
The privileges shown in the table are as seen in VMware vSphere 5.5.
Datastore -> Allocate space
Datastore -> Browse datastore
Datastore -> Low level file operations
Datastore -> Remove file
Datastore -> Rename datastore
Extension -> Register extension
Extension -> Unregister extension
Extension -> Update extension
Folder -> Delete folder
Folder -> Rename folder
Global -> Disable methods
Global -> Enable methods
Global -> Licenses
Host -> Configuration -> Maintenance
Host -> Inventory -> Add standalone host
Network -> Assign network
Resource -> Assign virtual machine to resource pool
Resource -> Remove resource pool
Resource -> Rename resource pool
Sessions -> Validate session vApp -> Delete vApp -> Rename vApp -> Add virtual machine
Virtual machine -> Snapshot Management -> Revert to snapshot
Virtual machine -> Configuration *
Virtual machine -> Interaction -> Answer question
Virtual machine -> Interaction -> Power Off
Data Protector (10.00) Page 336 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Virtual machine -> Interaction -> Power On
Virtual machine -> Inventory -> Create new
Virtual machine -> Inventory -> Register
Virtual machine -> Inventory -> Remove
Virtual machine -> Inventory -> Unregister
Virtual machine -> Provisioning *
Virtual machine -> Snapshot Management -> Create snapshot
Virtual machine -> Snapshot Management -> Remove snapshot l
Web service : Optionally, change the web service entry point URI. Default: /sdk
If you select Integrated security, which is only available for VMware vCenter Server systems with both the application client and the backup host are Windows systems, Data Protector connects to the VMware vCenter Server system with the user account under which the
Data
Protector Inet service on the backup system is running. Ensure that this user account has appropriate VMware vSphere rights to connect to the VMware vCenter Server system and the
Data Protector Inet service on the backup host is configured for user impersonation.
For details on setting accounts for the Inet service user impersonation, see the HPE Data
Protector Help index: “Inet user impersonation”.
For the Port and Web service root options, Data Protector uses the values that are currently specified for the standard security. Integrated security is based on Security Support Provider
Interface (SSPI).
Click Finish.
Importing a VMware vCenter Server client (login credentials)
Data Protector (10.00) Page 337 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
NOTE:
and Checking the configuration of VMware clients, on page 340
.
Changing the configuration of VMware clients
When you update the credentials for connecting to a VMware client (vCenter Server, ESX(i) Server client), you actually update the cell_info file which resides on the Data Protector Cell Manager.
Therefore, you can only change the login credentials if you have the Data Protector Clients configuration user right. For details on the Data Protector user rights, see the HPE Data Protector
Help index: ”user groups”.
To update the credentials, use the Data Protector GUI or CLI.
Using the Data Protector GUI
You can update the credentials in two different places: in the Clients or in the Backup context.
Clients context
1. In the Context List, click Clients.
2. In the Scoping Pane, expand Clients, and then select the client for which you want to change the login credentials.
3. In the Results Area, click the Login tab.
4. Update the credentials and click Apply.
Backup context
It is assumed that a backup specification for the VMware client for which you want to change the login credentials already exists.
1. In the Context List, click Backup.
2. Open a backup specification for the VMware client for which you want to change the login credentials.
3. In the Source page, right-click the client at the top, and select Configure.
4. In the Configure Virtual Environment dialog box, update the values and click OK.
Data Protector (10.00) Page 338 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Changing the configuration of a VMware vCenter Server or VMware ESX(i) Server client
Using the Data Protector CLI
1. Log in to the backup host, open the command prompt and change to the directory in which the vepa_util.exe
command is located.
2. Execute:
For Integrated security: vepa_util.exe
command
--config
--virtual-environment vmware
--host
VMwareClient
--security-model 1
For Standard security:
VMware vCenter Server or VMware ESX(i) Server client vepa_util.exe
command
--config
--virtual-environment vmware
--host
VMwareClient
Data Protector (10.00) Page 339 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
--security-model 0
--username
Username
{--password Password | --encoded-password Password}
[--webroot
WebServiceRoot]
[--port
WebServicePort]
The message
*RETVAL*0
indicates successful configuration.
For descriptions of the options, see the vepa_util.exe
man page or the HPE Data Protector
Command Line Interface Reference.
Checking the configuration of VMware clients
During the configuration check, Data Protector tries to connect to a VMware client using the login credentials from the cell_info
file on the Data Protector Cell Manager.
To verify the connection, use the Data Protector GUI or CLI.
Using the Data Protector GUI
You can verify the connection to a VMware client after you have created at least one backup specification for this client.
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications and then Virtual Environment. Click the backup specification for the VMware client to be checked.
3. In the Source page, right-click the VMware client and select Check configuration.
Using the Data Protector CLI
1. Log in to the backup host, open the command prompt and change to the directory in which the vepa_util.exe
command is located.
2. Execute:
VMware vCenter Server or VMware ESX(i) Server client vepa_util.exe
command
--check-config
--virtual-environment vmware
--host
VMwareClient
The message
*RETVAL*0
indicates successful configuration.
For option description, see the vepa_util.exe
man page or the HPE Data Protector Command
Line Interface Reference.
Configuring virtual machines
To configure a virtual machine means to specify how the virtual machine should be backed up.
You can specify the following:
Data Protector (10.00) Page 340 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
(Windows virtual machines only) Whether a quiescence snapshot should be taken to make applications running inside a virtual machine consistent for backup.
l
Which transportation mode should be used during backups.
For each datacenter, you can specify: l
Common settings that apply to all virtual machines in the datacenter.
l
Virtual machine-specific settings that override the common settings. If there are no virtual machinespecific settings, the common settings are used for that particular virtual machine.
All these settings are saved in a datacenter-specific file VMwareClient%DatacenterPath on the Cell
Manager. The file is used for all backup sessions that use any of the backup specifications for this datacenter.
Similarly, backup sessions using any of the backup specifications for All datacenters , use the settings from the file VMwareClient%AllDatacenters .
The files VMwareClient%DatacenterPath and VMwareClient%AllDatacenters are created or updated when you create or update a backup specification for a specific datacenter or all datacenters respectively.
To configure virtual machines, use the Data Protector GUI or CLI.
Using the Data Protector GUI
You can configure virtual machines when you create or modify a backup specification. In the Source page of a backup specification, right-click the client system at the top or any of the virtual machines listed below, and select Configure Virtual Machines.
Data Protector (10.00) Page 341 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Configure virtual machines (settings for VMware vCenter Server client)
In the Configure Virtual Machines dialog box Settings page, specify the following settings:
Virtual machine settings
Options Available Description / Action
Select whether you want to specify common virtual machine settings (Common VM Settings) or settings for a specific virtual machine. Virtual machine specific settings override the common virtual machine settings.
Configure virtual machines
Use common settings for selected
VM
Available only if a virtual machine is selected.
Select this option if you want the common settings to apply to the selected virtual machine.
Data Protector (10.00) Page 342 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Options Available
Use default settings
Enable changed block tracking
Allow fallback to non-CBT backups
Use quiescence snapshots
Level of error
Description / Action
Default: selected
Available only if Common VM Settings is selected.
Select this option to set default values for the common virtual machine settings.
Default: selected
The VMware changed block tracking functionality is enabled for the selected virtual machines.
Default: selected and greyed out
Enabled only if Use changed block tracking is selected.
Select this option to continue backup in a non-CBT mode for successful backup of Data Protector.
For more information on Non-CBT, refer to Non-Changed Block Tracking
(Non-CBT) backup, on page 320 topic.
Default: not selected.
Snapshot handling
Applicable for Windows virtual machines.
Not available if Use default settings or Use common settings for
selected VM are selected.
Select this option to use Microsoft Volume Shadow Copy Service (VSS) functionality to quiesce all applications with VSS writers before performing the VEPA backup. This produces application-consistent backups.
Default: not selected. For details, see Quiescence
.
Available only if Use quiescence snapshots is selected.
Specify the level of error to be reported if a quiescence snapshot fails.
Default: Warning. For details, see Quiescence .
Transportation mode
Select the transportation mode to be used when backing up a virtual machine.
l
NBD : Use this mode when your ESX(i) Server systems do not have access to a SAN, but use local storage devices or NAS to store virtual machine disks. This is an unencrypted transportation mode over a
Data Protector (10.00) Page 343 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Options Available Description / Action
local area network that uses the Network Block Device (NBD) driver protocol. This transportation mode is usually slower than Fibre
Channel.
l
NBD (SSL) : Same as NBD except that the communication over the network is encrypted using the Secure Socket Layer (SSL) cryptographic protocol.
l
Hotadd : Use this mode if your backup host (a client with the Data
Protector Virtual Environment Integration component installed) is a virtual machine. Such a configuration enables you to back up other virtual machines residing on datastores visible to the ESX(i)
Server that hosts the backup host.
l
SAN : Use this mode when your ESX(i) Server systems store their virtual machine disks on Fibre Channel SAN or iSCSI SAN. This is an unencrypted transportation mode over Fibre Channel or iSCSI.
o
Caution: Do not reformat these storage volumes. Otherwise, you will delete all your virtual machines.
For details on these VMware transportation modes, see the VMware documentation.
If you are not concerned which mode is used, select Fastest
available.
Default: Fastest available
Using the Data Protector CLI
1. Log in to the backup host, open the command prompt and change to the directory in which the vepa_util.exe
command is located.
2. Execute: vepa_util.exe
command
--configvm
--virtual-environment { vmware | vCD }
--host
AppHostName
--instance
DatacenterPath
--vm
VMpathVM_OPTIONSVM_OPTIONS
--transportation—mode {san | nbd | nbdssl | hotadd | fastest}
--quiescence { 0 | 1 }
--quiescenceErrLvl { 0 | 1 }
--uuid
UUID_of_VM
For details, see the vepa_util.exe
man page or the HPE Data Protector Command Line
Interface Reference.
To change the virtual machine specific settings back to the common virtual machine settings, execute:
Data Protector (10.00) Page 344 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware vepa_util.exe
command
--configvm
--virtual-environment { vmware | vCD }
--host
AppHostName
--instance
DatacenterPath
--vm
VMpath
--uuid
UUID_of_VM
--default
The message *RETVAL*0 indicates successful configuration.
Example
To configure the virtual machine with the virtual machine path /MyDatacenter/MyVM and UUID
42375365-ebe1-e9da-7068-7beb727cab19 that resides in the datacenter /MyDatacenter and registered in the vCenter Server system vc.company.com
, with the fastest available transportation mode.
execute: vepa_util.exe
command
--configvm
--virtual-environment vmware
--host vc.company.com
--instance
--vm
/MyDatacenter
/MyDatacenter/MyVM
--quiescence 1
--quiescenceErrLvl 0
--transportation-mode fastest
--uuid 42375365-ebe1-e9da-7068-7beb727cab19
Customizing the Data Protector behavior with omnirc options
The omnirc options are useful for troubleshooting or overriding other settings affecting the behavior of the Data Protector client. The options that apply to the Virtual Environment ZDB integration for VMware
have the prefix
OB2_VEAGENT
.
For details on how to use Data Protector omnirc options, see the HPE Data Protector Help index:
“omnirc options”.
Adding the RMC Server details in Data Protector using the Command Line Interface
You can add the RMC server details using the omnidb command. Execute: omnidb -addhost -servername <clientname> -user <username> -psswd <password>
Data Protector (10.00) Page 345 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
You can also use the following commands to list and remove the required servers. Execute: omnidb -listhost -servername <clientname> omnidb -removehost -servername <clientname> -user <username>
Backup
This section contains procedures that are needed to perform zero downtime backup of virtual
machines. For details on backup concepts, see Backup concepts, on page 314
.
With the Virtual Environment ZDB integration for VMware, you can perform the zero downtime backups for: l
NetApp storage: ZDB to tape l
3PAR storage: ZDB to disk, ZDB to tape, and ZDB to disk+tape
For more information on the ZDB types, see the HPE Data Protector Concepts Guide.
You should be familiar with the NetApp and HPE 3PAR storage concepts and procedures and basic
Data Protector ZDB functionality. See the storage-related documentation and the HPE Data Protector
Zero Downtime Backup Administrator's Guide.
For information about the supported configurations, ZDB types and replication techniques available on this storage system, and storage system-specific ZDB considerations, see the HPE Data Protector
Concepts Guide.
Backup limitations
l
Before performing backups,ensure that you use only the supported characters in the names of any
VMware vCenter and VMware ESX(i) objects (for example, virtual machines, datastores, datacenters, vApps, and so on), as special characters are not supported.
The following list includes the supported characters: o
Letters a-z without any special characters o
Numbers 0-9 o
Single Quotes (') o
Spaces o
Underscores (_) o
Hyphens (–)
A partial list of special characters that are not supported is as follows: o
% (percentage) o
+ (plus) o
= (equals)
Data Protector (10.00) Page 346 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware o
@ (at) o
< (less than) o
> (greater than) o
: (colon) o
" (double quote) o
/ (forward slash) o
\ (backslash) o
| (vertical bar or pipe) o
?
(question mark) o
* (asterisk)
Special characters are not supported due to a known VMware issue described at the following
URLs: http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd;=displayKC&externalId;=2017661 http://www.vmware.com/support/developer/vddk/VDDK-1.2.1-RelnoteOPs.html
l
After you upgrade, you cannot restart the failed backup object versions from earlier Data Protector versions below 8.11. l
After you upgrade from Data Protector 7.03 or earlier versions, you cannot run incremental or differential backups without running a full backup.
l
After an upgrade to Data Protector 9.05, the SAN transport mode falls back to NBDSSL on the vSphere versions 5.1 and 5.5. This is a VMware VDDK 6.0 Update1 issue. For more details, see the following URL: http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2135621 . l
For RMC integrations, backups that have virtual machines that are presented through iSCSI to the
ESXi server is not supported.
l
In RMC integrations, for Snapshot + Tape backups, the name of the configured mount ESXi host in the 3PAR array must be the same as its hostname. Also, the configured name of the mount ESXi in the 3PAR array must be in lower case. l
For datastores not located on VMFS volume, the full backups take full size of the disk and not only the changed blocks. For more details, see the following URL: http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=1020128 l
Consider all limitations that apply to the Data Protector NetApp Storage and HPE 3PAR Storage integration. The following limitations apply to Virtual Environment ZDB integration for VMware: o
ZDB backup is supported on NetApp storage and HPE 3PAR storage system. Multiple storage providers are not supported in a single backup session.
Data Protector (10.00) Page 347 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware o
You cannot perform ZDB backup on virtual machines or virtual disks that do not reside on a
NetApp storage or 3PAR storage.
o
A template backup in the SAN transport mode is not supported, and will fall back to the NDB transport mode.
o
If you want to back up a single disk from a virtual machine that contains multiple disks from different data stores, you must select the disk that belongs to the data store containing the configuration files.
o
Only one snapshot type for target volumes can be created during a ZDB session.
o
When cloning process for a source volume is in progress, another snapshot (any type) of that source volume cannot be created.
o
You cannot back up replicas (target volumes from existing and currently recorded backup sessions).
o
If there is not enough space for a fully allocated replica creation, the session fails.
o
Do not enable the vCenter option disable-datastore-web. This may result in unsuccessful backups of vmdk and vmx files. For more information, see vSphere Hardening Guides .
vStorage Image + OpenStack backup method limitations
l
ZDB backup and restore are not supported.
l
Non-CBT backups are not supported.
Creating backup specifications
Create a backup specification using the Data Protector GUI. The ZDB backup steps for NetApp storage and HPE 3PAR storage systems are similar.
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click Virtual Environment, and select Add Backup.
3. In the Create New Backup dialog box, select Snapshot or split mirror backup as Backup type and Storage Provider Plugin as Sub type. For description of options, press F1. Click OK.
4. Specify the application to be backed up: l
In the Client drop-down list, select a VMware client.
NOTE:
The drop-down list contains all clients that have been imported into the Data Protector cell as VMware vCenter or VMware ESX(i) clients. They have a corresponding label appended at the end of their names, such as (VMware vCenter) or (VMware ESX(i)).
Data Protector (10.00) Page 348 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
If the selected VMware client is not configured correctly, a warning is displayed. Click OK to open the Configure Virtual Environment dialog box and provide the connection parameters as
described in Importing and configuring VMware clients, on page 335
.
l
In the Backup host drop-down list, select a VMware vCenter system to be used to control the backup. The list contains all clients that have the Data Protector related ZDB integration component ( NetApp Storage Provider or 3PAR Storage Provider ) and Virtual
Environment Integration component installed.
l
In Datacenter/Organization, select a datacenter to back up from.
NOTE:
If you have selected a standalone ESX(i) Server system in the Client option, there is only one datacenter available – /ha-datacenter.
If you have selected a vCenter Server system in the Client option, you can select All
Datacenters to back up virtual machines from different datacenters.
l
In Mount host, select the ESX(i) Server system to be used for mounting the replicas.
NOTE:
If the mount proxy host with ESXi version 5.5 U2 belongs to the source data center,
Remote Copy failover backup will fail. Choose the ESXi mount proxy host from a different data center that belongs to the same vCenter.
l
In Backup method, the backup method is displayed: o vStorage Image for VMware vCenter and VMware ESX(i) clients o vStorage Image + OpenStack for VMware vCenter l
In Free space required [%], specify the percentage of disk space that should be free on a datastore before a virtual machine is backed up. The free space is calculated based on the size of the datastore where the virtual machine disks reside.
.
Data Protector (10.00) Page 349 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Selecting a VMware vCenter Server client, backup host, mount host, and datacenter
Click Next.
Selecting a Replication mode
Click OK.
Data Protector (10.00) Page 350 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
5. Under Add Storage Provider, select the desired storage provider from the Storage Provider dropdown list and then click Add. The storage provider specific options dialog opens. Select the replica provision type, enter a replica description, and specify other storage provider specific options, if available: l
Replica provision type: Thin provisioned or Fully allocated replica provision type are available for NetApp and 3PAR storage providers.
l
Transportation mode: SAN and NBD transportation modes are available for NetApp storage provider.
Click OK. The storage provider is added to a list. You can later change its options by clicking Edit or remove it from the list by clicking Remove. For more information, press F1. Click Next.
6. Under Replication mode, the HPE 3PAR Local Copy is selected by default. You can select the
HPE 3PAR Remote Copy as well. For more information, see the HPE Zero Downtime Backup
Administrator's Guide.
7. For 3PAR storage systems, select the Keep the replica after the backup check box. Choose the required value for Number of replicas rotated. Optionally, select the Track the replica for
instant recovery check box.
8. Select the objects that you want to back up. In the Show drop-down list, simplify your selection by choosing the Hosts and Clusters or VMs and Templates view. By default, Hosts and Clusters is displayed.
NOTE:
If you switch the view after you have already selected one or more objects for backup, a warning dialog is displayed. Its confirmation clears the already selected objects, clicking
No does no change to the view.
You can make your selections at different levels: l
For VMware vCenter and VMware ESX(i) clients: o
ESX/ESXi Servers systems o
Pools o vApps o
VM folders o
Individual VMs o
VM disks o
VM templates
If you select any level above individual VMs (for example, a vApp), all VMs and VM disks contained in the selected item will be included in the backup specification. If VMs are added within the item after the backup specification is saved, they will also be backed up.
NOTE:
If you clear the check box against an object, that object will be excluded from the backup specification. Thereafter, if you add a new virtual machine, a pool, or a vApp to an existing logical object, it is automatically included in the backup. You are not required to create a new backup specification. The check box for excluded objects is marked with a red cross
Data Protector (10.00) Page 351 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware symbol in the selected view.
Selecting VMware objects (vCenter Server client)
NOTE:
During the vStorage Image + Openstack backup specification creation, the Shadow VMs are not available for selection during backup and restore operations.
IMPORTANT:
In the object tree of a particular VMware client, a virtual machine may be displayed as selected in two different ways: l
The blue check mark indicates that the virtual machine is selected for backup in its entirety, including its configuration and all its virtual disks.
If such a virtual machine is backed up, it can be restored even if the original virtual machine does not exist anymore.
l
The gray or black check mark indicates that some or all virtual disks belonging to the virtual machine are selected. The virtual machine itself and its configuration is omitted from the backup.
If such a virtual machine is backed up, its disks can only be restored if the original virtual machine is still configured at the time of restore.
If your virtual machines are not configured yet, right-click the client system at the top or any of the virtual machines listed below, and select Configure Virtual Machines. For details, see
Configuring virtual machines, on page 340 .
Click Next.
9. Select which devices to use for the backup.
To specify device options, right-click the device and select Properties. Specify the number of parallel backup streams in the Concurrency tab and which media pool to use.
Click Next.
Data Protector (10.00) Page 352 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
10. Set backup options.
.
Application-specific options
Click Next.
12. Click Save As to save the backup specification, specify a name and a backup specification group. Optionally, you can click Save and Schedule to save, and schedule the backup specification. For more information on how to create and edit schedules, see Scheduler in Data
Protector in HPE Data Protector Administrator's Guide.
13. Click Start Backup to start the backup session. If you are using the 3PAR storage system, you will have the option of selecting the Backup type, Network load, and the Snapshot backup options.
Data Protector (10.00) Page 353 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
TIP:
.
VMware backup options
Option Description
Pre-exec , Post-exec Specifies which command line to run on the backup host before ( preexec ) or after ( post-exec ) the backup.
Do not use double quotes. Type only the name of the command and ensure that the command resides in the default Data Protector administrative commands directory on the backup host.
Windows systems: Data_Protector_home\bin
Linux systems: /opt/omni/bin
Creating backup specifications for RMC backups
Consider all the prerequisites and limitations listed in the StoreOnce Recovery Manager Central
section.
Create a backup specification using the Data Protector GUI.
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, right-click Virtual Environment, and select Add Backup.
3. In the Create New Backup dialog box, select StoreOnce RMC as Backup type. Select
Snapshot, Snapshot+Tape, or Express Protect as as Sub type, based on your requirements.
Click OK.
l
Snapshot: If you select this, replicas of the source volumes are created.
l
Snapshot+Tape: If you select this, replicas of the source volumes are created, and the data is backed up to the chosen tape device. Ensure that you specify the Mount Host, and the required Devices in the subsequent steps. l
Express Protect: If you select this, you can backup snapshots from HPE 3PAR StoreServ to
HPE StoreOnce. The backup proceeds as per the backup policy created in RMC.
Data Protector (10.00) Page 354 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
4. Specify the application to be backed up: l
In the Client drop-down list, select a VMware client. If the selected VMware client is not configured correctly, a warning is displayed. Click OK to open the Configure Virtual
.
l
In the Backup host drop-down list, select a VMware vCenter system to be used to control the backup.
l
In Datacenter/Organization, select a datacenter to back up from.
l
In Mount host, select the ESX(i) Server system to be used for mounting the replicas. This field is mandatory for Snapshot+Tape backups.
l
In Backup method, select the method to be used for the backup.
NOTE:
RMC does not support the vStorage Image + OpenStack backup method.
5. Specify the RMC configuration details:
Data Protector (10.00) Page 355 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
RMC Server: Select the RMC server for backup. l
Backup Policy (available only for Express Protect backups): Select the policy name that you have created in RMC. The backup policy usually contains the backup system and the backup store. For more information, see the HPE StoreOnce RMC User Guide. l
Snapshot Count: Specify the maximum number of snapshot count that should be maintained in the storage array. A maximum of 1000 and a minimum of 1 can be created. The default value is 10. l
Express Protect Count (available only for Express Protect backups): Specify the maximum number of backups that should be maintained in StoreOnce for the backup specification. The minimum value that can be set is 2. The default value is 10.
6. Select the objects that you want to back up. In the Show drop-down list, simplify your selection by choosing the Hosts and Clusters or VMs and Templates view. By default, Hosts and Clusters is displayed.
7. Select the devices that you want to use for the backup. If you have specified the mount host, then the device is selected by default.
Click Next.
NOTE: This step is not applicable for Snapshot and Express Protect backups.
8. Set backup options. Click Next.
9. Click Save As to save the backup specification, specify a name and a backup specification group. Optionally, you can click Save and Schedule to save, and then schedule the backup
Data Protector (10.00) Page 356 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware specification. For more information on how to create and edit schedules, see Scheduler in Data
Protector in HPE Data Protector Administrator's Guide.
10. Click Start Backup to start the backup session.
Modifying backup specifications
To modify your backup specification, click its name in the Scoping Pane of the Backup context, then click the appropriate tab, and apply the changes.
In the Source page, you can modify the backup objects by using the Show drop-down list. In the dropdown list, the Hosts and Clusters view or the VMs and Templates view are available. The view that you have used during the creation of your backup specification has the string (Original) appended. If you switch the view, a warning dialog is displayed and its confirmation clears the already selected objects.
NOTE:
When modifying a backup specification created with one of the previous versions of Data
Protector, you can switch between Selected, All, Hosts and Clusters, and VMs and
Templates views. The view All is available for the legacy backup specifications only and provides a legacy browsing mechanism. Selecting Hosts and Clusters or VMs and
Templates and clicking Yes in the warning dialog clears the previous backup object selection and upgrades your browsing mechanism. After saving the backup specification only the Hosts
and Clusters and VMs and Templates views are available.
Modifying a backup specification (VMware vCenter Server client)
To display the virtual environment settings, in the Results Area, click VE Settings. Not all settings can
be modified.
Data Protector (10.00) Page 357 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Scheduling backup sessions
You can run unattended backups at specific times or periodically. For more information on how to create and edit schedules, see Scheduler in Data Protector in HPE Data Protector Administrator's
Guide.
Previewing backup sessions
Preview the backup session to test it. You can use the Data Protector GUI or CLI.
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications and then Virtual Environment. Right-click the backup specification you want to preview and select Preview Backup.
3. Specify the Backup type and Network load. Click OK.
The message Session completed successfully is displayed at the end of a successful preview.
Using the Data Protector CLI
1. Log in to any client with the Data Protector User Interface
component installed.
2. Open the command prompt and change to the directory in which the omnib command is located.
3. Execute: omnib -veagent_list BackupSpecificationName -test_bar
What happens during the preview?
The following are tested: l
Communication between the backup host and Data Protector l
The syntax of the backup specification l
If devices are correctly specified l
If the necessary media are in the devices
Starting backup sessions
Interactive backups are run on demand. They are useful for performing urgent backups or restarting failed backups.
To start a backup interactively, use the Data Protector GUI or CLI.
Data Protector (10.00) Page 358 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Using the Data Protector GUI
1. In the Context List, click Backup.
2. In the Scoping Pane, expand Backup Specifications, and then Virtual Environment. Rightclick the backup specification you want to use and select Start Backup.
3. Specify Backup type and Network load. Click OK.
The message Session completed successfully is displayed at the end of a successful backup session.
Using the Data Protector CLI
1. Log in to any client with the Data Protector User Interface component installed.
2. Open the command prompt and change to the directory in which the omnib command is located.
3. Execute: omnib -veagent_list BackupSpecificationName [-barmode VirtualEnvironmentMode]
[ListOptions ] where
VirtualEnvironmentMode
is one of the following backup types: full|diff|incr
The default is full .
For
ListOptions
, see the omnib man page or the HPE Data Protector Command Line Interface
Reference.
Examples
To start a full backup using the backup specification MyVirtualMachines , execute: omnib -veagent_list MyVirtualMachines -barmode full
To start a differential backup using the same backup specification, execute: omnib -veagent_list MyVirtualMachines -barmode diff
Preparing for disaster recovery
To do a disaster recovery, you need backups of the following VMware objects:
What must be backed up
VMware object How to back it up
ESX/ESXi Server console
ESX Server systems:
1. Ensure that the Data Protector Disk Agent component is installed on the backup system.
2. In the Backup context of the Data Protector GUI, right-click
Filesystem and select Add backup to create a backup specification of the filesystem type. In the Source page of the backup specification, select ESX Server consoles of all ESX
Data Protector (10.00) Page 359 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
VMware object
vCenter configuration database
(applicable only for
VirtualCenter environments)
VMware virtual machines
How to back it up
Server systems.
For details on what to back up, see the topic “ESX Server
Configuration Backup and Restore procedure” at http://kb.vmware.com/selfservice/microsites/microsite.do
.
3. Start a backup using the newly created backup specification.
ESXi Server systems:
With ESXi Server systems, it is not possible to install a Data
Protector Disk Agent , so you will need to use VMware utilities to back up the configuration.
A tool, esxcfg-cfgbackup is available from VMware. For information, see the VMware website.
The vCenter configuration database can be an Oracle database or
Microsoft SQL Server database. To back up the database, use the corresponding Data Protector integration. For example, if it is an Oracle database, follow these steps:
1. Ensure that the Data Protector Oracle Integration component is installed on the vCenter Server system.
2. In the Backup context of the Data Protector GUI, right-click Oracle
Server and select Add backup to create a backup specification of the Oracle type. In Application database, type the name of the vCenter configuration database.
Continue with the backup specification creation as described in the
HPE Data Protector Integration Guide.
3. Start a backup using the newly created backup specification.
Back up the virtual machines as described in this chapter.
Restore
This section contains procedures that are needed to restore virtual machines. For details on restore
concepts, see Restore concepts, on page 326
.
Restore limitations
l
When restoring a VM to a datacenter with a datastore shared among several inventory objects
(hosts or clusters) or several cluster nodes, and the restore option for its subsequent registration is selected, the restored VM may not register at the original inventory location:
Data Protector (10.00) Page 360 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware o
If several inventory objects share the datastore, the VM is registered with the first available host or cluster.
o
If several cluster nodes share the datastore, the VM is registered with the first available cluster node.
If the VM must be registered at its original inventory location, migrate it appropriately after the restore session completes.
l
Before performing a restore,ensure that you use only the supported characters in the names of any
VMware vCenter, or VMware ESX(i), objects (for example, virtual machines, datastores, datacenters, vApps, and so on), as special characters are not supported. For more details on the supported characters, see Backup Limitations.
l vDatacenters that are disabled do not allow adding or removing of vApps. Consequently restore to a disabled vDatacenter is not supported.
l
In a vSphere environment, do not use the left parenthesis (the
(
character) in names of vSphere distributed port groups. Data Protector is unable to connect a restored virtual machine to a nonoriginal vSphere network that uses distributed switches and is assigned such a distributed port group. In this case, after the restore, you need to manually connect the restored virtual machine to the desired network using vSphere Client.
l
Recovering virtual machines after a restore to a directory is not supported for incremental or differential backups. The VMware Converter that is used for moving the restored VM images to an
ESX Server or ESXi Server system recognizes only the full backup type.
l
Recovering virtual machines from a backup chain (for example, full, incremental, incremental, incremental...) is supported only for a restore to a datacenter.
l
All the media related to a session should be exported or imported together.If a media from the full backup object is missing then the Incremental/Differential backup will not detect a missing media and continue to run in the selected mode. A restore from such a session will not be successful.
l
A virtual machine can be seen as different objects in the Internal Database (IDB) and in the restore context based on how the backup specification was created.
l
It is not possible to restore a suspended Virtual Machine.
l
In earlier versions of Data Protector, all disks are read even if you select only one disk for restore.
From Data Protector 8.11 onwards, only the selected disks are read and restored. This is applicable for backups created with Data Protector 8.11 and not with its earlier versions.
l
Movement from one disk to another on a tape is not possible. If you select consecutive disks
(scsi0:1 and scsi0:2) for restore, then both the disks are read and restored. If you select scsi0:1 and scsi0:4 for restore then the following combinations occur: o scsi0:0 is read.
o scsi0:1 is read and restored.
o scsi0:2 is read.
o scsi0:3 is read.
o scsi0:4 is read and restored.
Data Protector (10.00) Page 361 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
So, random selection of disks could result in higher restore window when compared to single or consecutive disk selection.
l
During a VEPA (Virtualization Environment Agent) restore process, the Virtual Environment
Integration Agent makes a connection to the vCenter / ESX to upload the virtual machine configuration files to VC / ESX. Such a connection could continue to exist in vCenter even after the restore session is complete. This idle session is cleared by VMware after a default time-out value of
30 minutes. Alternatively this session can be cleared manually from the vCenter.
vStorage Image + OpenStack restore limitations
l
Restore from Replica is not supported.
l
Features, such as, Restore-as, Delete-after, and Forensics are removed from the Data Protector
Restore context. l
Granular Recovery Extension is not supported for Nova Instance and Shadow VM.
l
If the Instances are deleted from the OpenStack dashboard, Recovery of instance from the Data
Protector to Dashboard is not possible.
l
If the Shadow VM is attached to any other instance other than the original instance to which it was attached at the time of backup, then the restore fails.
l
Restore to a different cluster is not supported.
NOTE:
The Power On and Live Migrate functionalities are not supported for vStorage Image +
OpenStack method.
Power On and Live Migrate limitations
l
The Power On and Live Migrate options are not supported for Smart Cache device that are created on CIFS and NFS shares. l
Power On and Live Migrate operations from Storeonce catalyst are supported only with Linux backup hosts.
l
For virtual machines that are created on Smart Cache and StoreOnce Catalyst devices, only eight virtual machines can be powered on at once. If you want to increase this number, see Increasing the default value that defines the maximum number of NFS mounts on an ESXi/ESX host . l
If you have backed up a virtual machine using the Disk Only option, restore to an ESX server is not supported. l
If you have backed up a virtual machine using the Disk + Tape option, restore to an ESX server is supported from the tape. l
If you Power On a virtual machine from Smart Cache and StoreOnce Catalyst devices, the virtual machine will be powered on without the virtual RDM disks.
l
If you perform a Live Migrate procedure from Smart Cache and StoreOnce Catalyst devices, the virtual RDM disks will be migrated as thick disks, i.e. after migration they will reside as eager zeroed thick provisioned vmdk disks on the target datastore. If you want to restore virtual machines that have virtual RDM disks, use the normal Restore procedure.
l
Power On and Live Migrate of virtual machines to ESX servers outside the vCenter, is not
Data Protector (10.00) Page 362 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware supported.
l
Live Migrate of backups that are performed with an ESX server as the backup client, is supported only through the CLI. l
You can perform only one operation (Restore / GRE / Power On / Live Migrate) at a time on the virtual machines that have backups on StoreOnce Catalyst device. l
If the virtual machine has ongoing restore operations (Power On, Live Migrate, or GRE) from the
StoreOnce Catalyst device, clean up the virtual machine before performing any of the restore operations from the same backup session. This is also applicable for the backup sessions that are part of the restore chain. l
StoreOnce Catalyst devices does not support Power On and Live Migrate options for Non-CBT backup sessions. l
Power On and Live Migrate operations from the StoreOnce Catalyst devices are not supported if the selected backup host is RHEL 6.6.
l
Power On and Live Migrate operations are not supported if the backup to StoreOnce Catalyst device is performed using software compression or AES encryption.
l
If Power On and Live Migrate of a virtual machine is performed by selecting vApp as destination, then the virtual machine will be powered on and live migrated outside the vApp. You can manually move it inside the vApp if required.
l
If the mount proxy host is rebooted with any active Power On or Live Migrate request, then this request will be inaccessible for 4 hours after the reboot.
l
If the data backed up with Data Protector 9.04 or earlier versions is transferred to a StoreOnce
Catalyst through object copy in 9.07 or later version, then Live Migrate and Power On operations are not supported.
l
Full, differential, and incremental backups on different devices are not supported for Power On and
Live Migrate operations.
For example:
Full backups of the same virtual machine are done to the StoreOnce Catalyst and Smart Cache devices respectively.
Incremental backup is performed to the StoreOnce Catalyst device. This incremental backup is not eligible for Power On and Live Migrate operations because the full backup preference for this incremental chain resides on Smart Cache.
Finding information for restore
You can find information about backup objects in the Data Protector IDB, such as which backup type and media were used, and which messages were displayed during the backup. To retrieve this information, use the Data Protector GUI or CLI.
Using the Data Protector GUI
In the Internal Database context, expand Objects or Sessions.
If you expand Objects, backup objects are sorted according to the virtual machine for which they were created. For example:
Data Protector (10.00) Page 363 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
In a vCenter environment, the backup objects for the virtual machine /vm/mach1 are listed under
/4/vCenterName%2FvmInstanceUUID
.
Where, vCenterName is the name of the virtual center.
vmInstanceUUID is a unique identifier for virtual machine /vm/mach1 on vCenter.
To view sessions in the vCenter environment, double-click the
/object1/vCenter%2FvmInstanceUUID
.
VEAgent Object details
If you expand Sessions, backup objects are sorted according to the session in which they were created. For example, backup objects created in the session 2012/07/10–82 are listed under
2012/07/10–82 .
To view details on a backup object, right-click the backup object and select Properties.
Data Protector (10.00) Page 364 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Backup object information
TIP:
To view the messages displayed during the session, click the Messages tab.
Using the Data Protector CLI
1. Log in to any client with the Data Protector User Interface component installed.
2. Open the command prompt and change to the directory in which the omnidb
command is located.
3. Get a list of VMware backup objects created in a backup session with the session ID SessionID: omnidb -session SessionID
4. Get details on a backup object with the backup object name BackupObjectName: omnidb -veagent BackupObjectName -session SessionID -catalog
Here is one example of a backup object name: gabriel.company.com::/%2FElDatacentro/0/%2Fvm%2Fharbour
For details, see the omnidb man page or the HPE Data Protector Command Line Interface Reference.
Restoring using the Data Protector GUI
Use this procedure to Restore, Power On and Live Migrate virtual machines.
1. In the Context List, click Restore.
2. In the Scoping Pane, expand Virtual Environment, expand the relevant client and click the datacenter from which you backed up.
3. In the Source page, specify the following:
Data Protector (10.00) Page 365 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
a. From the Backup method drop-down list, select any of the following backup methods: l vStorage Image for VMware vCenter and VMware ESX(i) clients l vCD vStorage Image for VMware vCloud Director clients l vStorage Image + OpenStack for VMware vCenter in OpenStack environment
b. From the From and To drop-down lists, you can narrow the scope of displayed virtual machines to those backed up within the specified time interval.
c. In the VM Filter text box, enter filter text for the VM and press Enter , or click Apply Filter .
The filter hides the VMs, vApps, and resource pools that do not match the filter pattern allowing you to find the desired object easily.
After you select the VMware objects, you can choose to Restore, Power On or Live Migrate them. Choose the required option from the VM Options drop down. Note that the Power On and
Live Migrate options are available only when one object is selected.
NOTE:
Filters are case-sensitive and apply to VMware Virtual Machine objects, VMware Virtual
Application (vApp) objects, and resource pools. If a matching sub-node is found (such as, another VM, vApp, or resource pool), they are displayed. If you leave the VM Filter text box empty, all the VMs, vApps, and resource pools are displayed. However, if you enter filter text, only the matching sub-nodes (if any) are displayed. If you modify the filter values using the From , or To drop-down lists, filtering is re-applied. Filter does not apply to already selected VMs, vApps, or resource pools. This means that you can filter machines using one filter, select objects, and again change the filter. In the new filter, the previously marked objects remain visible.
The following are the types of filter usage: l
Simple substring usage - If you enter part of a VM, vApp, or resource pool name, all VMs, vApps, or resource pools in the object tree that have the entered string in their name remain visible. All other objects are filtered out. l
Wildcards and question mark usage - The following are the filtering options: o
<filter string>* - Filters VM, vApp, or resource pool names that start with the <filter string> and end with any set of characters.
o
*<filter string>*
- Filters VM, vApp, or resource pool names that start and end with any set of characters, and have the <filter string> in between.
o
*<filter string>
- Filters VM, vApp, or resource pool names that start with any set of characters and end with the
<filter string>
.
o
<filter string>*01 - Filters VM, vApp, or resource pool names that start with the
<filter string>
, end with a "01", and have any set of characters in place of the wildcards
(*). For example,
Production_VM01
.
o
<filter string>*0?
- Filters VM, vApp, or resource pool names that start with the
<filter string>
, end with a "0" and have a character, number, or letter in place of the question mark (?). For example,
Production_VM01
and
Production_VM0A
, but not
Production_VM11 .
Select the objects that you want to restore.
Data Protector (10.00) Page 366 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
NOTE:
Data Protector restores the full restore chain for each selected VMware object, beginning with the last full backup session (even if that full backup is outside the specified time interval) and ending with the last backup session performed during the specified time interval.
Selecting VMware objects for restore (vCenter Server client)
Power On ( PowerOnOption ON ): Power on the virtual machine from the backed up image, Smart
Cache, or StoreOnce Catalyst device.
Live Migrate ( PowerOnOption MIGRATE ): Live Migration of the virtual machine from the backed up image, Smart Cache, or StoreOnce Catalyst device.
Data Protector (10.00) Page 367 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Selecting vStorage Image + OpenStack backup method for restore
Restoring as a new virtual machine
Right-click the selected virtual machine and then click Restore As / Into to restore it as a new virtual machine for vStorage Image backup method. A new dialog box opens to specify the name of the virtual machine.
NOTE:
The Restore As/Into option is specific to vStorage Image backup method.
When you select the vStorage Image + OpenStack backup method, you get to view the Nova instances that were backed up and its versions. The Shadow VM objects are not displayed during restore.
Data Protector (10.00) Page 368 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Restoring selected backup version
Right-click the selected virtual machine and click Restore Version to select a backup version that you want to restore.
A new dialog box opens for selecting the backup version. The object versions for the selected
Nova Instances are displayed.
NOTE:
(Specific to vStorage Image backup method) A warning message
VM is being restored to a different data center is displayed in the dialog box when the backup version is selected from a different datacenter.
Data Protector (10.00) Page 369 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
4. In the Destination page, specify the restore destination. For details, see Restore destination
(VMware vCenter Server and VMware ESX(i) Server clients), on page 372
.
Data Protector (10.00) Page 370 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Restore destination (VMware vCenter Server client)
5. If you have selected Power On or Live Migrate, click the respective button to complete the operation.
.
Restore options (VMware vCenter Server and VMware ESX(i) Server clients)
Data Protector (10.00) Page 371 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Restore options (vStorage Image + OpenStack)
Here, VMs are deleted before restore, and Nova Instance and Shadow VMs are registered in the vCenter after restore.
7. In the Devices page, select devices to use for restore.
8. Click Restore.
9. In the Start Restore Session dialog box, click Next.
10. Specify Report level and Network load.
NOTE:
Select Display statistical information to view the restore profile messages in the session output.
11. Click Finish to start the restore.
The statistics of the restore session, along with the message Session completed successfully
is displayed at the end of the session output.
NOTE:
If the restore fails, see Cleaning up a datastore after a failed restore, on page 387 .
RMC integration restore
l
For RMC backups, if the backup type is Snapshot+Tape, preference is given for the restore operation from the snapshot. If the snapshot is absent, the restore happens from the tape device.
l
If the backup type is Express Protect, restore happens to the 3PAR LUNs first, and subsequently, to the location specified.
Restore destination (VMware vCenter Server and VMware ESX(i) Server clients)
Data Protector (10.00) Page 372 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
GUI/CLI option
Backup host /
-barhost
Restore client /
-apphost
Restore to datacenter /
–instance – newinstance
Host/Cluster /
-host/cluster
Specific host /
-specificHost
Resource pool /
-resourcePool
Datastore /
–store
Description
Specifies the client with the Virtual Environment ZDB Integration for
VMware installed to control the restore session. By default, the same client as used for the backup is selected.
Specifies the client that the selected virtual machine objects should be registered and restored to. By default, the client from which the virtual machines were backed up is selected.
To change the client configuration, click the Connect button.
Select this option to restore a virtual machine to a datacenter. By default, virtual machines are restored to the original datacenter.
You can restore from a 3PAR array to a datacenter.
Select the ESX(i) Server system or the cluster to which virtual machines should be restored. By default, virtual machines are restored to the original ESX(i) Server system or cluster.
Select the specific ESX(i) Server system in the cluster to which virtual machines should be restored. By default, virtual machines are restored to the original ESX(i) Server system.
Select the resource pool on the ESX(i) Server system or the cluster to which virtual machines should be restored. By default, virtual machines are restored to the original resource pool.
Specifies the datastore to which the virtual machines should be restored. You can choose among all datastores that are accessible from the selected restore target host. If you leave this option empty, the virtual machines are restored to the original datastore.
Restore to directory /
–directory
Select this option to restore virtual-machine files to a directory
(outside of the datacenter) on the backup host. You can use the
Browse button to find the target directory.
After such a restore, the virtual machines are not functional. You need to manually move the restored virtual machine images to an ESX
Server or ESXi Server system, using the VMware Converter as
described in Recovering virtual machines after restore to a directory, on page 380
.
Restore options
GUI/CLI option
Register virtual machines if needed /
Description
Available if Restore to datacenter
Data Protector (10.00) Page 373 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
GUI/CLI option
-register
Consolidate snapshots to single file /
-consolidate
Power-on virtual machines after restore /
-poweron
Description
is selected.
Select this option to register restored virtual machines.
If this option is not selected, you need to manually recover the restored virtual machines as
described in Recovering virtual machines manually, on page 380
.
Default: selected.
NOTE:
The Register virtual
machines if needed option is unavailable for selection with vStorage Image +
OpenStack backup method.
Select this option to commit all snapshots (including non-Data
Protector ones) to the virtual machine base once a virtual machine is restored.
Available if Restore to datacenter is selected.
Select this option to power the virtual machines on once they are restored.
Available if Restore to datacenter is selected.
Data Protector (10.00) Page 374 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
GUI/CLI option
Existing virtual machine handling
NOTE:
This option is unavailable for selection with
vStorage Image +
OpenStack backup method. The Delete
before restore option is used.
Description
Specifies Data Protector's behavior when restoring existing virtual machines.
Data Protector (10.00) Page 375 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
GUI/CLI option
Delete before restore /
-deletebefore
Skip restore /
-skip
Delete after restore /
-deleteafter
Keep for forensics /
–keep_for_forensics
Description
Select this option to delete an existing virtual machine before it is restored, and then restore it from new. The existing virtual machine is deleted even if it resides in a different datacenter from your target datacenter.
This is the space-efficient option, but is less secure, since the old virtual machine is not available if the restore fails, so select it with caution.
Select this option to skip the restore of an existing virtual machine. This allows you to restore missing virtual machines without affecting existing ones.
Select this option to delete an existing virtual machine after it is restored. The existing virtual machine is deleted even if it resides in a different datacenter than your target datacenter. If the restore fails, the existing virtual machine is not deleted.
Default: selected.
This option is not available for
VMware vCloud Director client.
This option cannot be used if the virtual machine is in a suspended state.If the virtual machine is in a suspended state, do any of the following: l
Restore to a different location.
l
Select the Delete before
restore option.
l
Power On or Off the virtual machine.
Select this option to mark an existing virtual machine with a
Data Protector (10.00) Page 376 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
GUI/CLI option
File conflict handling
NOTE:
This option is unavailable for selection with
vStorage Image +
OpenStack backup method. The
Overwrite option is used.
Description
timestamp. The virtual machine which is kept for forensics is powered off after the restore and remains at the original location. It does not affect consecutive backups of the original virtual machine.
This option is not available for
VMware vCloud Director and
Microsoft Hyper-V clients.
Specifies Data Protector's behavior when restoring existing files.
Overwrite /
-overwrite
Select this option to overwrite existing files with those from the backup.
Default: selected.
Keep latest /
-latest
Skip /
-skip
Select this option to leave an existing file intact if it is more recent than the one from the backup.
Otherwise, the existing file is overwritten with the one from the backup.
Select this option to preserve an existing file (the file is not restored from the backup).
Restoring using the Data Protector CLI
1. Log in to any client with the Data Protector User Interface component installed.
2. Open the command prompt and change to the directory in which the omnir
command is located.
3. Execute:
VMware vCenter Server or VMware ESX(i) Server client omnir -veagent
-virtual–environment vmware
-barhost
BackupHost
-apphost
OriginalVMwareClient
–instance
OriginalDatacenter
–method vStorageImage| vStorageImageOpenStack
Data Protector (10.00) Page 377 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
[-session
[
BackupID]
VirtualMachine [VirtualMachine...]
VMwareClient | Directory]
VirtualMachine
-vm VMPath -instanceUUID vmInstanceUUID [-new_name disk
DiskName [-disk Disk...]]
NewVirtualMachineName][-
VMwareClient
[–newinstance
[-store
TargetDatacenter]
TargetDatastore]
[-network_name
TargetNetwork]
[–destination
TargetVMwareClient]
[-consolidate] [–register][–poweron]
[–deletebefore | -deleteafter | -skip | -keep_for_forensics]
Directory
-directory
RestoreDirectory
[–overwrite | -skip | —latest]
Restore Options
[-consolidate] [-register] [-poweron] [-PowerOnOption { ON | MIGRATE }] [deletebefore | -skip | -keep_for_forensics]
For a description of all the options, see the omnir man page or the HPE Data Protector Command
Line Interface Reference.
NOTE:
You should not specify the instanceUUID parameter for restore while restoring the virtual machine backed up from Data Protector 8.1 and below.
IMPORTANT:
A backup ID is a point in time. All objects (backup data) created in a backup session have the same backup ID, which is the same as the session ID of the backup session.
Mirrored objects and objects created in an object copy session have the same backup ID as objects created in the original backup session. Suppose the media set created in the original backup session no longer exists, but the media set created in an object copy session still exists. To restore the objects, you have to specify the session ID of the original backup session (that is, the backup ID) and not the session ID of the object copy session.
The omnir
syntax does not let you specify from which object copy to restore if several copies of the same object exist. This is only possible using the Data Protector GUI by setting the media allocation priority list.
NOTE:
If the restore fails, see Cleaning up a datastore after a failed restore, on page 387 .
Data Protector (10.00) Page 378 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Example (Restoring virtual machines to a datacenter)
Suppose you want to restore the virtual machine
/vm/machineA
and the individual disks ( scsi0:0
and scsi0:1
) of the virtual machine
/vm/machineB
. At the time of backup, the virtual machines were running on the ESX Server systems that belonged to the datacenter /MyDatacenter managed by the vCenter Server system vcenter.company.com
. The virtual machines were backed up with the vStorage Image backup method.
To restore them to the original location, using the backup session 2011/01/11–1 , and to ensure that the newly restored virtual machines are put online when the session completes, execute: omnir -veagent -virtual–environment vmware -barhost backuphost.company.com -apphost vcenter.company.com –instance /MyDatacenter -method vStorageImage –session
2011/1/11-1 –vm /vm/machineA –vm /vm/machineB -disk scsi0:0 –disk scsi0:1 –poweron
To restore the virtual machine /vm/machineA with an instanceUUID 503eeaac-6fae-7898-73e1-
93b722a0517c , execute: omnir -veagent -virtual–environment vmware -barhost backuphost.company.com -apphost vcenter.company.com –instance /MyDatacenter -method vStorageImage –session
2011/1/11-1 –vm /vm/machineA -instanceUUID 503eeaac-6fae-7898-73e1-93b722a0517c disk scsi0:0 –disk scsi0:1 –poweron
Example (Restoring virtual machines to a directory)
Suppose the virtual machines /MyVirtualMachines/machineA and /MyVirtualMachines/machineB were backed up in the session
2011/02/12-5
from the datacenter
/MyDatacenter
that is managed by the vCenter Server system vcenter.company.com
, using the vStorage Image
backup method. To restore the virtual machines outside of the datacenter, to the directory C:\tmp on the backup host backuphost.company.com
, execute: omnir -veagent -virtual–environment vmware -barhost backuphost.company.com -apphost vcenter.company.com -instance /MyDatacenter -method vStorageImage -session
2011/2/12-5 -vm /MyVirtualMachines/machineA -vm /MyVirtualMachines/machineB directory c:\tmp
Example (Restoring object names with instanceUUID in its name)
To support restore of object names with instanceUUID in its name, execute: omnir.exe -veagent -virtual-environment vmware -barhost barHostName -apphost appHostName -instance instanceName -method vStorageImage -session sessionID -vm vmPath –instanceUUID vminstanceUUID -register -poweron -deletebefore
Example (Restoring Nova Instance to its original location)
To restore object names with instanceUUID in its name, execute: omnir -veagent -virtual-environment vmware -barhost barHostName -apphost appHostName -instance /Datacenter -method vStorageImageOpenStack -session sessionID
-vm vmPath -instanceUUID novainstanceUUID -register -deletebefore
Data Protector (10.00) Page 379 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Recovering virtual machines manually
There are two different scenarios in which you need to recover virtual machines manually after they have been restored with Data Protector: l
If you have restored the virtual machines to a directory on a backup host (Restore to directory).
For details, see Recovering virtual machines after restore to a directory, below .
l
If you have restored the virtual machines to a datacenter (Restore to datacenter) without selecting the restore option Register virtual machines if needed.
For details, see Recovering virtual machines after restore to a datacenter, on page 386 .
Recovering virtual machines after restore to a directory
The steps for recovering virtual machines after restore to a directory depend on the format in which the virtual machine configuration file was backed up.
Recovering with the VM configuration file in the VMX format
Suppose the virtual machine helios was restored to the directory C:\tmp\helios on the backup host using the following backup session: l
Backup method: vStorage Image l
Backup type: Incremental l
CBT: Enabled and used
To move the virtual machine files manually to the ESX(i) Server system dioxide.company.com
managed by the vCenter Server system bmwvc2.company.com
, using the VMware Converter:
1. Display the contents of the directory C:\tmp\helios : helios.vmdk
helios.vmx
helios.vmdk
scsi0–0.cbt
scsi0–0.meta
helios-flat.vmdk
helios.vmx-1 helios.vmdk-1 scsi0–0.cbt-1 scsi0–0.meta-1 helios.vmx-2 helios.vmdk-2 scsi0–0.cbt-2 scsi0–0.meta-2
Note that all files backed up in the last full, differential, and the selected incremental session are restored.
2. Share the folder C:\tmp\helios so that it can be accessed from the system with the VMware
Converter installed.
Data Protector (10.00) Page 380 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
3. Log in to the system with the VMware Converter installed and open the VMware Converter user interface.
4. Click Convert Machine to open the Conversion wizard.
5. In the Source System page, select VMware Workstation or other VMware virtual machine for the source type, browse to the C:\tmp\helios directory, and select the helios.vmx
file.
Click Next.
Conversion (Source System)
NOTE:
In our example, the VMware Converter is installed on the backup host.
6. In the Destination System page, select VMware Infrastructure virtual machine for the destination type and provide the login credentials for the vCenter Server system.
Click Next.
Data Protector (10.00) Page 381 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Conversion (Destination System)
7. In the Destination Virtual Machine page, specify the name under which the virtual machine should be recovered.
Click Next.
Conversion (Destination Virtual Machine)
8. In the Destination Location page, select the destination ESX(i) Server system and datastore.
Data Protector (10.00) Page 382 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Conversion (Destination Location)
9. In the Options page, edit the options and click Next.
Data Protector (10.00) Page 383 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Conversion (Options)
10. In the Summary page, review your selection and click Finish.
Data Protector (10.00) Page 384 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Conversion (Summary)
11. Open the Datastore Browser and upload the files created in the incremental and differential backup sessions to the virtual machine directory: helios.vmx-1 helios.vmdk-1 scsi0–0.cbt-1 scsi0–0.meta-1 helios.vmx-2 helios.vmdk-2 scsi0–0.cbt-2 scsi0–0.meta-2
12. Power the virtual machine on.
Data Protector (10.00) Page 385 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Recovering with the VM configuration file in the XML format
Follow the procedure:
1. Open vSphere Client and log in to an ESX(i) Server or vCenter Server system.
If the virtual machine is still configured, remove all its hard disks:
a. In the inventory object tree, right-click the virtual machine and select Edit Settings.
b. In the Virtual Machine Properties window, in the Hardware tab, select each hard disk and click Remove.
c. Click OK to confirm the removal.
If the virtual machine is no longer present, configure a new virtual machine without hard disks, and use the name of the original virtual machine.
In either case, remember the associated datastore name.
2. Upload the virtual machine files that were created during the backup session:
a. In the inventory objects tree, select the ESX(i) Server system that hosts the virtual machine.
b. Click the Configuration tab and select Storage under Hardware.
c. Right-click the datastore name and select Browse Datastore.
d. In the Datastore Browser window, in the folder tree, select the virtual machine folder, and click a corresponding icon on the window toolbar. Select Upload File or Upload Folder as appropriate.
e. Select all applicable files and complete the upload.
3. Add hard disks to the virtual machine by reusing their backup copies:
a. In the inventory object tree, right-click the virtual machine and select Edit Settings.
b. In the Virtual Machine Properties window, click Add.
c. In the Add Hardware window, select Hard Disk and click Next.
d. Select Use an existing virtual disk and click Next.
e. Click Browse.
f. In the Browse Datastores window, browse to the appropriate datastore and open the virtual machine folder. Select the virtual disk file and click OK.
g. Follow the Add Hardware wizard to complete the process.
h. Repeat the substeps from b to g for each additional hard disk for which a backup copy exists.
4. Power the virtual machine on.
Recovering virtual machines after restore to a datacenter
If you have restored a virtual machine to a datacenter without selecting the option Register virtual machines if needed:
1. Open the Datastore Browser and browse to the restored virtual machine directory.
2. Right-click the virtual machine *.vmx
file and select Add to Inventory.
3. Follow the wizard and click Finish.
Data Protector (10.00) Page 386 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Restoring using another device
You can restore using a device other than that used for backup. For details, see the HPE Data
Protector Help index: “restore, selecting devices for”.
Cleaning up a datastore after a failed restore
Sometimes, when a virtual machine restore fails, Data Protector creates extra files on the virtual machine datastore. If these files are not deleted, corrupt virtual machine backups may be created in subsequent sessions and, consequently, restore from such a backup also fails.
Suppose the virtual machine MyVirtualMachine failed to be restored. To clean up the datastore after the restore:
1. Open the VMware vSphere client.
2. Right-click the virtual machine and select Delete from disk.
3. Open the Datastore Browser.
The directory
MyVirtualMachine
should no longer be there.
Check if there are any extra directories:
MyVirtualMachine_1
MyVirtualMachine_2 and so on.
Right-click each such directory and select Delete from disk.
Disaster recovery
Disaster recovery is very complex, involving different products from different vendors. Check the instructions from the guest operating systems and VMware on how to prepare for it.
The following are the main steps needed to recover a virtual machine after a disaster:
1. Reinstall the VMware environment. The configuration should be the same as during the backup.
2. Install Data Protector in the newly configured environment.
3. Restore the service console of the ESX Server system on which the virtual machine was running to the newly configured ESX Server system from a Data Protector filesystem backup.
For details on what to restore, see the topic “ESX Server Configuration Backup and Restore procedure” at http://kb.vmware.com/selfservice/microsites/microsite.do
.
For details on how to restore from a filesystem backup, see the HPE Data Protector Help.
4. Restore the original vCenter database (if needed). For details, see the Data Protector integration that was used to back up the database.
5. Restore the virtual machine from a Data Protector Virtual Environment backup as described in this chapter.
Data Protector (10.00) Page 387 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Instant recovery
For general information on instant recovery, see the HPE Data Protector Concepts Guide and the HPE
Data Protector Zero Downtime Backup Administrator's Guide.
Prerequisites
l
ESX Server l
3PAR arrays l vCenter l
Application host on virtual machine l
Backup host on physical machine
NOTE:
Both the hosts including the ESX Server should have access to 3PAR arrays.
HPE 3PAR ZDB Instant recovery
Data Protector offers HPE 3PAR ZDB instant recovery for agents within a VMware Virtual Machine
(VM). This is supported only for physical Raw device mapping (RDM).
To perform an instant recovery on Linux and Windows:
1. Shut down the Oracle database instance using sqlplus . In case of RAC, shut down all instances.
For example:
/sqlplus /nolog connect sys/oracle@APPN as sysdba sql> shutdown immediate sql> exit
2. Enable the omnirc option on Application host:
ZDB_IR_MANUAL_AS_PREPARATION=1
3. For Linux: Dismount volume:
# umount /dev/3PAR_ESX2/lvol0
For Windows: Offline the disks.
4. Prepare Application Host to remove volume (export, deactivate, and backup Volume Group).
5. Remove hard disks from Application host on vCenter server.
6. Rescan volumes on VM, and confirm that the disk is not presented any more on the Application
Host.
7. Execute Instant Recovery.
NOTE:
If you are using the Oracle integration, ensure that you deselect the Recovery checkbox.
8. Add the hard disks back to the application host from the vCenter server.
9. Rescan Application Host for new volume.
Data Protector (10.00) Page 388 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
10. Add the exported Volume Group.
11. For Linux: Mount volume.
For Windows: Online the disks.
Monitoring sessions
You can monitor currently running sessions in the Data Protector GUI. When you run a backup or restore session, a monitor window shows the progress of the session. Closing the GUI does not affect the session.
You can also monitor sessions from any Data Protector client with the User Interface component installed, using the Monitor context.
To monitor a session, see the HPE Data Protector Help index: “viewing currently running sessions”.
Troubleshooting
This section lists general checks and verifications, plus problems you might encounter when using the
Data Protector Virtual Environment ZDB integration for VMware.
For general Data Protector troubleshooting information, see the HPE Data Protector Troubleshooting
Guide.
Before you begin
l
Ensure that the latest official Data Protector patches are installed. For more information, see the
HPE Data Protector Help index: “patches”.
l
See the HPE Data Protector Product Announcements, Software Notes, and References for general
Data Protector limitations, as well as recognized issues and workarounds.
l
See https://softwaresupport.hpe.com/ for an up-to-date list of supported versions, platforms, and other information.
Checks and verifications
If your configuration, backup, or restore failed: l
Examine system errors reported in the debug.log
on the backup host.
l
Check if you can do a filesystem backup and restore on the problematic client. For information, see the HPE Data Protector Help.
Additionally, if your backup failed: l
Check the configuration of the vCenter Server or standalone ESX(i) Server system as described in
Checking the configuration of VMware clients, on page 340
.
Additional actions:
Data Protector (10.00) Page 389 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
After performing the file delete operation in a guest VM, the full CBT backup session has the similar backed up data as the previous full CBT backup session. The reason is that the VMware does not reclaim the disk space after the file delete operation is performed in the guest VM.
Workaround-1:
1. Reclaim the unused space of a virtual disk in ESXi/ESX 4.1 or later. For details, see http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2004155
2. Disable and re-enable the CBT. For details, see http://kb.vmware.com/selfservice/search.do?cmd=displayKC&docType=kc&docTypeID=DT_
KB_1_1&externalId=1031873
Workaround-2:
Perform a Storage vMotion on the virtual machine or VMDK to a datastore formatted with a different block size. For example, if the VMDK is on a datastore formatted with 2 MB blocks, format the target
VMFS datastore with a 1 MB, 4 MB, or 8 MB block size. Once this is done, the CBT has to be reset.
l
Before enabling CBT and running a backup, ensure that user-created snapshots are not available in the virtual machine. Otherwise, the following error message is displayed: cannot be enabled.
For details, see http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=1020128
Workaround: o
Delete all user-created snapshots associated with the virtual machine before enabling CBT.
o
Consolidate changes, if the changes in snapshot needs to be backed up using CBT.
l
An incremental or differential CBT backup session falls back to a FULL CBT backup session when performing an incremental or differential CBT backup session after a virtual disk migration using the
VMware Storage VMotion, the backup falls back to a FULL CBT session.
Workaround:
The issue is resolved with ESX 5.5 update 2. The reason is that the CBT is reset after a virtual disk migration using the storage VMotion. For details, see http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2048201
Problems
Problem
An incremental or differential CBT backup session fails
When performing an incremental or differential backup session with the Use changed block tracking option enabled, the session fails with an error similar to the following:
Data Protector (10.00) Page 390 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
[Critical] From: [email protected] "/New Datacenter 4.1"
Time: 2/10/2011 11:14:52 AM
Virtual Machine 'ddd': Could not gather changed blocks on disk scsi0:0 ...
The reason may be that you performed a restore session and forgot to run a full backup session to start a new backup chain.
To make sure changed block tracking is working properly, certain requirements need to be met. For details, see http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd;=displayKC&externalId;=1020128
Action
1. Run a full backup session.
2. Run an incremental or differential backup session.
Problem
After a restore or move to a different folder, backups are not performed correctly
After you have restored or moved a virtual machine to a different folder, the virtual machine is not backed up correctly. For example, instead of an incremental backup, a full backup is performed.
The reason for this is that the datacenter configuration file has been updated. As a result, it contains two virtual machine sections with the same UUID; this is from where the inconsistencies arise.
Action
Re-configure the virtual machine:
1. Open the backup specification.
2. In the Source page, right-click the VMware client and select Configure Virtual Machines.
3. Click OK.
If addition of disks is done after the backup on the backed up Nova instance, the following error message is displayed:
Error: Virtual Machine '9d28cd95-c158-45ec-b606-53f7c63a2a78': Cannot perform
Restore. Found new disk attached to the Nova Instance.
When a user needs a point in time restore, and a new disk was added after that point in the backup, restore is not allowed. User has to detach the new volumes attached to the instance from the
OpenStack and perform the restore again.
It is required to perform a full backup after the addition of the new disks.
Problem
A restore session using SAN transportation mode fails
A restore session that is using SAN transportation mode fails with a message similar to the following:
[Critical] From: [email protected]
"/BlrVirtual01_ESX401" Time: 13-03-2011 12:22:57
Virtual Machine 'Win2k3_x64_dpi00002': Error restoring item
Data Protector (10.00) Page 391 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
\a1f9f4e3-482d-4b7f-afcb-cb16babe1980\%2FBlrVirtual01_ESX401\vms
\%2FBlrVirtual01_ESX401%2Fhost%2FClus01%2FWin2k3_x64_dpi00002\ images\3\scsi2:15.
This may happen if a storage volume that is shared between the backup host and an ESX(i) Server system is read-only.
Action
1. Log in to the backup host and open the command prompt.
2. Execute diskpart .
C:\Users\Administrator>diskpart
Microsoft DiskPart version 6.1.7600
Copyright (C) 1999-2008 Microsoft Corporation.
On computer: TPC134
3. Set the SAN policy to onlineAll .
DISKPART> san policy=onlineAll
DiskPart successfully changed the SAN policy for the current operating system.
4. Select the disk (storage volume) that should be used for restore.
DISKPART> list disk
Disk ### Status Size Free Dyn Gpt
-------- ------------- ------- ------- --- ---
Disk 0 Online 136 GB 1024 KB
Disk 1 Offline 14 GB 14 GB
Disk 2 Offline 14 GB 14 GB
Disk 3 Offline 14 GB 14 GB
Disk 4 Offline 14 GB 14 GB
Disk 5 Offline 50 GB 50 GB
Disk 6 Offline 14 GB 14 GB
Disk 7 Offline 14 GB 14 GB
DISKPART> select disk 1
5. Bring the disk online.
DISKPART> online disk
DiskPart successfully onlined the selected disk.
6. Ensure the disk is not read-only.
List the disk properties.
DISKPART> detail disk
Data Protector (10.00) Page 392 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
HP OPEN-V SCSI Disk Device
Disk ID: 00000000
Type : FIBRE
Status : Online
Path : 0
Target : 0
LUN ID : 0
Location Path : UNAVAILABLE
Current Read-only State : Yes
Read-only : Yes
Boot Disk : No
Pagefile Disk : No
Hibernation File Disk : No
Crashdump Disk : No
Clustered Disk : No
There are no volumes.
Clear the read-only attribute.
DISKPART> attribute disk clear readonly
Disk attributes cleared successfully.
List the disk properties again.
DISKPART> detail disk
HP OPEN-V SCSI Disk Device
Disk ID: 00000000
Type : FIBRE
Status : Online
Path : 0
Target : 0
LUN ID : 0
Location Path : UNAVAILABLE
Current Read-only State : No
Read-only : No
Boot Disk : No
Pagefile Disk : No
Hibernation File Disk : No
Crashdump Disk : No
Clustered Disk : No
There are no volumes.
Exit the session.
DISKPART> exit
7. Restart the restore session.
Data Protector (10.00) Page 393 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Problem
Slow performance of vepa_util.exe browse command on newer Red Hat Enterprise Linux
(RHEL) versions
When executing the vepa_util.exe browse command on a newer RHEL version its performance is significantly slower than on other operating systems.
Action
The root cause of the problem is that on newer versions of RHEL systems name service cache daemon is not enabled by default.
Start the name service cache daemon by invoking the following command: /etc/init.d/nscd start .
To enable automatic daemon start-up during system start-up, execute: chkconfig nscd on .
Problem
A restore job fails when restoring virtual machines to ESX(i) hosts managed by vCenter
Server 5.x or higher
When restoring a virtual machine to an ESX(i) host, which is managed by a vCenter Server 5.x or higher, the restore job fails and the virtual machine is not restored successfully. Starting with ESX(i)
5.0, VMware has blocked the ability to restore a virtual machine to an ESX(i) which is managed by vCenter.
Action
To resolve this issue, either restore virtual machines through the vCenter Server or restore them through ESX/ESX(i) after unmanaging the host from the vCenter Server.
For more details on how to unmanage an ESX/ESXi host from vCenter Server, see http://kb.vmware.com/kb/2038838
Problem
Restore of a virtual machine using an ESX(i) Server system ends with a corrupted VM guest operating system
In a vSphere environment, when you restore a virtual machine to a
/ha-datacenter
using an ESX(i)
5.0 Server system or later as a restore client, the restore session ends successfully, but the guest operating system on the virtual machine is corrupted.
Action
When selecting the objects for restore, in the Destination page, in the Restore client drop-down list, select a vCenter Server instead of an ESX(i) Server system.
Problem
In some cases, when backup is started and the backup specification contains multiple Virtual
Machines, the following message appears:
[Major] From: [email protected] "backup_spec_name" Time: 4/10/2014 2:55:07 PM
[61:2052] Bar backup session was started but no client connected in 600 seconds.
Aborting session!
Data Protector (10.00) Page 394 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
The VEAgent collects metadata information for all the Virtual Machines before starting the backup process. While the VEAgent is busy collecting information, the BSM waits for 10 minutes (default timeout) for the VEAgent to get connected. After the default time has elapsed the BSM times-out with the message shown above.
Action
Increase the value of the timeout variable available in Data Protector Global Options.
SmWaitForFirstBackupClient=WaitForInMinutes
SmWaitForFirstBackupClientSec=WaitForInSeconds
Problem
Backup of Virtual Environments can fail with the following error messages: l
Exception occurred while creating VM snapshot l
Backup of object failed
These errors may occur if the timeout values on vCenter and ESX hosts are low.
Action
Increase the timeout values in the vCenter configuration file ( vpxd.cfg
) and the ESX hosts configuration file ( vpxa.cfg
).
For more details on the resolution, see the following: http://kb.vmware.com/kb/1017253 http://kb.vmware.com/kb/1005757
NOTE: NOTE: You can change the values suggested here based on your requirements.
In addition, check the timeout values and operational timeout values in the vCenter:
Increase the operational timeout values on vCenter.
1. Log in to the vCenter server with administrator credentials.
2. Go to the following location: vCenter Server Settings > Timeout Settings.
3. Under Client Connection Timeout
, set the following values: l
Normal Operations timeout : 600 l
Long Operations timeout : 2000
Increasing the timeout value in vCenter
To increase the number of idle connections between ESX and the vCenter host, add the following values in the vpxd.cfg
file within the
<config>
and
</config>
tags:
<vpxd>
<maxHostPooledConnections>20000</maxHostPooledConnections>
</vpxd>
Data Protector (10.00) Page 395 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
NOTE: NOTE: This will reduce the creation of additional TCP connections to the proxy during host synchronization.
Problem
After upgrading, if the backup specification is created using Data Protector 7.00, 7.01, or earlier versions, the VEAgent backup fails with the following error:
[Critical] From: VEPALIB_VMWARE@<hostname> "<Datacenter>"
No Objects found for backup
Action
After upgrading, perform the following steps:
1. Re-create the backup specification with the same VM selection and options as it was earlier.
2. Run the backup again.
Problem
The Virtual Environment Integration agent (VEPA) and the Session Manager stall while waiting for the time-out value to elapse.
While running multiple parallel VEPA backup sessions, you could encounter a scenario where a small fraction (1 or 2) of the total number of sessions may stall in such a way that the VEPA and BSM are not responsive until the time-out period. This could be because the vCenter is loaded with multiple concurrent connection requests from the VEPA agent, from parallel VM backup sessions.
The stalled VEPA and BSM processes for these sessions may finally time-out with the following message:
[Major] From: BSM@machineName "barlist7" inactivity timeout of xxxx seconds. The agent on host will be shutdown .
Action
After the time-out period:
1. Stop the vepa_bar processes manually and wait for the BSM process associated with it to close as well.(The BSM closes after the vepa_bar
exits.)
2. Restart the backup specification that contains the failed VM objects.
Note: If the time-out period elapses before the backup starts (while resolving the objects), then increase the SmWaitForFirstBackupClient parameter in Internal Database -> Global Options.
Problem
Parallel backup sessions fail
If the VMs in the backup specification belong to the same LUN (already being backed up), and the mount proxy hosts that are used for the backups are the same then, for parallel backups the second backup session will fail.
Action
Use a different mount proxy host for the second backup session.
Data Protector (10.00) Page 396 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Problem
The zero downtime backup of virtual machines from the 3PAR replica fails with the following
error message:
If the backup specification has large number of virtual machines or the load on the virtual center or the
ESX server is high, the zero downtime backups may fail with following error message:
[Major] From: [email protected] "New2" Time: MM/DD/YYYY HH:MM:SS AM
[61:2052] Bar backup session was started but no client connected in XX seconds.
Aborting session!
Action
Extend the timeout value by resetting the Data Protector SmWaitForFirstBackupClient global options variable. For details on how to set the variable, see the Data Protector online help.
Problem
For Linux virtual machines, the IP configured is lost after performing a restore operation.
This is likely to happen in the following scenario:
Before performing a Restore operation, if you click Keep for forensics or Delete after restore option in the restore work flow, then the restored VM will have a new
NIC set to
DHCP
, and its original
NIC goes into hiding.
Action
This is a known VMware / Linux limitation. For a virtual machine that is running a Linux guest operating system, while restoring that virtual machine, the ESX server may assign the virtual machine a new
(virtual) MAC address. After restarting the virtual machine, you may have to configure its MAC address. For example, the original MAC address of the virtual machine may be in a configuration file that has to be updated as mentioned in the following KB articles.
http://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2002767 and http://www.uptimemadeeasy.com/vmware/fixing-eth0-mac-address-vmware-clone-restore/
Problem
Virtual machine restore: Could not find object on three disks in virtual environment.
While restoring a virtual machine, the object could not be found on the three disks in the virtual environment.
Action
To enable partial restore of a virtual machine that is no longer available on the data store, proceed as follows:
Data Protector (10.00) Page 397 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
1. Create a temporary virtual machine with the same UUID as the original backup.
2. Restore partial data from backup to the temporary virtual machine.
Problem
Data Protector Virtual Environment integration (VEPA) backup sessions that use Backup to
Disk (B2D) gateways may fail.
In virtual environment backups that use B2D gateways, when the Cell Manager, VEPA agent, Media
Agent gateway for the server side deduplication are on the same machine, the backups may fail.
Action
You may perform the following tasks to solve this problem: l
Move the Media Agent, and the VEPA agent to another host.
l
Reduce the number of concurrent streams in the Media Agent. l
Set the omnirc variable OB2BMAUPDT to 10000. l
Increase the process priority of the Backup Session Manager. l
Switch to source-side deduplication. l
Switch the VEPA backups from SAN more to NBD mode.
l
Avoid monitoring sessions in the Data Protector GUI.
Problem
Backups of VMware virtual machine disk may fail
Backups of VMware virtual machine disk may fail with the following error message:
[ 20] [VddkUtil::diskLibWarning] VixDiskLib: Failed to load vixDiskLibVim.dll :
ErrorCode = 0x7f.
Action
This is a known VMware issue, and may be solved if you perform the following tasks: l
Install the vCenter and the VDDK on separate servers.
l
Copy the libldap_r.dll
and liblber.dll
from \Program Files\OmniBack\lib\vddk\AMD64 to
\Program Files\OmniBack\bin.
For more details on the resolution, see http://kb.vmware.com/kb/2120818 .
Problem
Booting of the Windows virtual machine fails after restore
When the Windows virtual machine is booted after a restore operation, the booting process fails with the following error: "No operating system found" .
Action
This is a known VMware problem. This error may occur if the operating system disks and the data disks are from different controller types, such as SCSI and IDE. For more details, see http://kb.vmware.com/kb/1023592 .
Data Protector (10.00) Page 398 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
You must change the boot order of the virtual machine using the options bios.bootOrder
and bios.hddOrder
. For the detailed procedure, see http://kb.vmware.com/kb/2011654 .
Problem
VMware ZDB backup, Power On and Live Migrate may fail
Some times, the VMware ZDB backup session, Power On and Live Migrate sessions may fail with the following error message:
[Critical] From: VEPALIB_VMWARE@<HostName> "<AppName>" Time: <Timestamp>
Error mounting datastores
Action
This error may occur because the source ESX host or the mount ESX may contain some erroneous unresolved disk volumes that cannot be resolved to the data store. You must identify such volumes on the source ESX server or the mount proxy ESX server host, and remove the presentation of those volumes from the hosts.
Problem
Error message appears during Power On and Live Migrate
The following error message appears during Power On and Live Migrate sessions:
Failed to install/start NFS service.
Action
Open the Windows PowerShell window on the mount proxy host, and run the following command:
PS C:\Program Files\OmniBack\bin> .\nfsServiceCheck.ps1
True
PS C:\Program Files\OmniBack\bin>
Problem
Error while creating virtual machine snapshots and object backup fails
An error occurs when creating the snapshots of a virtual machine :
[Critical] From: VEPALIB_VMWARE@<HostName> "<AppName>" Time: <Timestamp>
Error mounting datastores
[Normal] From: VEPALIB_VMWARE@BACKUPHOSTNAME "/DATACENTERNAME"
Virtual Machine ‘VMNAME’: Creating snapshot ...
[Major] From: VEPALIB_VMWARE@BACKUPHOSTNAME "/DATACENTERNAME "
Virtual Machine ‘VMNAME’: Error removing snapshot
[Critical] From: VEPALIB_VMWARE@BACKUPHOSTNAME "/CPD2" Time: 19/03/2016 8:01:48
Backup of object failed.
Name: VMNAME
Data Protector (10.00) Page 399 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Path: / DATACENTERNAME /DATASTORE/ VMNAME
InstanceUUID: IUUIDOFVM
This issue occurs when the virtual machines are located in Site Recovery Manager (SRM).
Action
The create snapshot API is disabled for virtual machines in SRM and therefore, backup operation is not supported.
Problem
Virtual Machines with virtual hardware version 4 fails to boot after the restore to Data Center
When the Virtual Machine is restored to datacenter from the Data Protector GUI, the virtual machine fails to boot although the restore is successful.
Action
1. From the Data Protector GUI, restore virtual machine to a folder on to backup host.
2. From the vCenter, create a new virtual machine without any disks attached.
3. Go to the vCenter datastore browser, and upload the vmdk
files from the folder created in Step 1
.
4. Edit the VM Settings from the vCenter to attach the uploaded vmdk files.
5. Restart the virtual machine to boot.
Problem
GRE, Power On, and Live Migrate operation fails
GRE, PowerOn and Live Migrate operation fails with the following error message:
Object locked: The VM <VM Name> could be locked by another process for recovery/power on/live migrate.
Please retry after the process is either done or cancelled .
Action
Ensure the following: l
Recover Files window displaying the Browse option is closed. l
The virtual machine does not have ongoing restore operations (Object Copy, Restore, Power On,
Live Migrate, or GRE) from the StoreOnce Catalyst device.
l
Clean up the powered on virtual machine.
Problem
Restore cannot be performed, if a new disk found is attached to the Nova instance
If a new disk found is attached to the Nova instance, the following error message is displayed:
Virtual machine '9d28cd95-c158-45ec-b606-53f7c63a2a78': Cannot perform Restore.
Found new disk attached to the Nova instance.
Data Protector (10.00) Page 400 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Action
When a user needs a point in time restore, and a new disk was added after that point in the backup, restore is not allowed. User has to detach the new volumes attached to the instance from the
OpenStack and perform the restore again.
Problem
At the time of restore, if the Shadow VM part of the backed up Nova Instance is attached to another Nova instance, the following error message is displayed in the session logs:
Virtual machine '9d28cd95-c158-45ec-b606-53f7c63a2a78': Can not perform restore, as the related shadow VM is attached to another Nov Instance '8d28ad65-c158-45ec-b606-
53f7c63a4578
Action
Detach the Shadow VM from the new instance and start the Restore.
Problem
The restored instance in the OpenStack dashboard does not reflect the correct state and remains in error state
Action
After the restore completes, restart the Compute Nova proxy service, and manually reset the error
.
Problem
Restore or object operation might fail if GRE, Power On, or Live Migrate operation is being performed for the same VEPA backup session to StoreOnce Catalyst
The object operations (object copy) will fail for a backup session “X” if GRE is being performed for the same object which is backed up in session “X”. Following error is displayed in Data Protector object copy session:
Cannot open device (StoreOnce error: The object is already locked and multiple open sessions not supported, or the server is unable to lock any more objects due to resource constraints)
Action
Perform the following:
l
Ensure that the virtual machine does not have ongoing restore operations (Object Copy, Restore,
Power On, Live Migrate, or GRE) from the StoreOnce Catalyst device.
l
Clean up the powered on virtual machine.
l
Retry after the active GRE, Power On, or Live Migrate operation is completed.
Problem
A backup session to StoreOnce Catalyst is not eligible for Power On and Live Migrate if:
Data Protector (10.00) Page 401 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware l
"No Logs" option is selected in the backup specification or l
The following error message is seen in the backup session report: information for device <Catalyst_device> with loaded medium <Catalyst_medium> to Data Protector Internal Database:
There is no more space available in any of the Detail Catalog directories.
From this point on, all objects on this medium will have logging switched to "No
Log".
Action
The following actions can be taken: l
Remove the "No Logs" option or l
Create a space on the IDB drive on the cell manager and perform a single session copy to another device. Ensure that replication option is not selected when single session copy is being performed to another StoreOnce Catalyst device.
Problem
Data consistency issues during Power On and Live Migrate operations
Data consistency issues during Power On and Live Migrate operations due to following reasons:
1. User has chosen a session which is a result of object copy.
2. Object copy is performed by aggregating many backup sessions into single object copy session as shown in the figure below:
Action
Choose individual session while performing object copy as shown in the figure below:
Data Protector (10.00) Page 402 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Problem
Restore of Linux virtual machine completes successfully, but ifconfig displays missing NIC
After performing restore of Linux virtual machine, the network adapters are given a new MAC address.
Then the ifconfig command running on restored virtual machine displays missing NIC.
For more information, see https://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2002767
Action
To details on updating the MAC address of the Linux operating system, see the Vmware KB article: https://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2002767
Problem
During backup of virtual machine, the following error message is displayed:
Virtual Machine <virtual_machine_name>: Download of disk descriptor file (<disk_ vmdk_file_name>) failed.
Action
Check whether the vpxd.enableHttpDatastoreAccess and vpxd.enableDebugBrowse are enabled in vCenter configuration file, as these settings might be disabled due to security reason.
For the detailed procedure, see https://kb.vmware.com/selfservice/microsites/search.do?language=en_
US&cmd=displayKC&externalId=2101567
Problem
Shadow VM is not attached to Nova Instance after restore
At the time of restore, if the backed up Shadow VM is not attached to the original backedup Nova instance, then the restored shadow VM is detached after the restore.
Data Protector (10.00) Page 403 of 413
Zero Downtime Backup Integration Guide
Chapter 6: Data Protector Virtual Environment ZDB integration for VMware
Action
To have the consistency with the Openstack dashboard, the restored shadow VM is detached during the restore process. To resolve, re-attach the shadow VM to Nova Instance from the openstack
Dashboard.
Data Protector (10.00) Page 404 of 413
Chapter A: Appendix
This appendix contains information on the following topics: l
Reconfiguring an Oracle instance for instant recovery l
ZDB integrations omnirc options
Reconfiguring an Oracle instance for instant recovery
If the control files or redo logs are located on the same volume group (if LVM is used) or source volume as the database files, the control files and online redo logs are overwritten during instant recovery. In such case, you may want to reconfigure the Oracle instance.
For details on the required configuration, see Oracle backup set ZDB concepts , on page 22
and to Oracle proxy-copy ZDB concepts , on page 26 .
Moving online redo logs
To move the online redo log files from the source volumes to be replicated, to other locations:
1. List the online redo log files:
$ sqlplus
SQL> select member from v$logfile;
2. Shut down the database:
SQL>connect user/password@service as sysdba;
SQL> shutdown
SQL> exit
3. Move the log files to a different location using operating system tools.
4. Start the database in mount mode:
$ sqlplus
SQL> connect user/password@service as sysdba;
SQL> startup mount;
5. Register the new locations for each moved file:
SQL> alter database rename file 'OldPathName' to 'NewPathName'; where OldPathName and NewPathName are full paths to the log file.
6. Open the database in normal mode:
SQL> alter database open;
Moving control files
To move the control files from the source volumes to be replicated, to other locations:
1. Determine if the database uses the SPFILE parameter:
Data Protector (10.00) Page 405 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
SQL> show parameter SPFILE
2. If the database does not use SPFILE:
a. Shut down the database.
SQL> shutdown
b. Move the control files to a different location using operating system tools.
c. Edit the CONTROL_FILES parameter in the database's initialization parameter file (usually located in the $ORACLE_HOME/dbs/initSID.ora
directory) to change the existing control file names: control_files = ("NewPathName", ...)
d. Restart the database:
SQL> startup
If the database uses SPFILE:
a. Specify the new location for control files by running the following command:
SQL> alter system set control_files='NewPathName1',
'NewPathName2',..., scope=spfile
b. Shut down the database.
SQL> shutdown
c. Move the control files to a different location.
d. Restart the database:
SQL> startup
Examples for moving the control files and redo logs to different locations
Example - moving online redo logs
In the following example for Oracle10g on HP-UX, the data files are on the same source volume as the control files and redo logs, which is /opt/oracle/product/10.2.0
.
To move the online redo log files from /opt/oracle/product/10.2.0
to /oracle/logs (which is not replicated):
1. List the online redo log files:
$ sqlplus
SQL> select member from v$logfile;
/opt/oracle/product/10.2.0/oradata/redo01.log
/opt/oracle/product/10.2.0/oradata/redo02.log
/opt/oracle/product/10.2.0/oradata/redo03.log
List the filenames and tablespaces to check whether they are on the same source volumes as the control files:
SQL> select FILE_NAME,TABLESPACE_NAME,BYTES from dba_data_files;
Data Protector (10.00) Page 406 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
FILE_NAME
---------------------------------------------------------
TABLESPACE_NAME BYTES
------------------------------ ----------
/opt/oracle/product/10.2.0/oradata/system01.dbf
SYSTEM 419430400
/opt/oracle/product/10.2.0/oradata/undotbs01.dbf
UNDOTBS1 377487360
/opt/oracle/product/10.2.0/oradata/cwmlite01.dbf
CWMLITE 20971520
2. Shut down the database:
SQL>connect user/password@service as sysdba;
SQL> shutdown
SQL> exit
3. Move the log files to a different location.
$ mv /opt/oracle/product/10.2.0/oradata/redo* /oracle/logs
4. Start the database in mount mode:
$ sqlplus
SQL> connect user/ password@service as sysdba;
SQL> startup mount;
5. Rename the new locations for each moved file:
alter database rename file '/opt/oracle/product/10.2.0/oradata/redo01.log' to
'/oracle/logs/redo01.log';
Database altered.
alter database rename file '/opt/oracle/product/10.2.0/oradata/redo02.log' to
'/oracle/logs/redo01.log';
Database altered.
alter database rename file '/opt/oracle/product/10.2.0/oradata/redo03.log' to
'/oracle/logs/redo01.log';
Database altered.
6. Open the database in normal mode:
SQL> alter database open;
Example - moving control files for Oracle 10g
In the following example, the Oracle 10g database uses SPFILE. To move the control files from
/opt/oracle/product/10.2.0/ to /oracle/oractl :
1. Determine if the database uses the SPFILE parameter:
SQL> show parameter spfile;
NAME TYPE VALUE
------------------------- ----------- ------------------spfile string ?/dbs/[email protected]
Data Protector (10.00) Page 407 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
2. Specify the new location for the control files by running the following command (in a single line and without the "\" characters):
SQL> alter system setcontrol_files='/oracle/logs/RCVCAT \
/control01.ctl','/oracle/logs/RCVCAT/control02','/oracle \
/logs/RCVCAT/control03.ctl' scope=spfile;
3. Shut down the database:
SQL>shutdown
4. Move the control files to the new location: mv /opt/oracle/product/10.2.0/oradata/control* /oracle/oractl
5. Restart the database:
SQL>startup
ZDB integrations omnirc options
The Data Protector ZDB integrations use omnirc options, which can be set on both the application and backup systems. These options are used for Data Protector ZDB integrations customizing. For information on how to use the omnirc options, see the HPE Data Protector Help index: “omnirc options”.
For information on Data Protector ZDB agents omnirc file options, see the HPE Data Protector Zero
Downtime Backup Administrator's Guide.
This section explains the omnirc file options that can be set for the Data Protector ZDB integrations.
ZDB_ORA_INCLUDE_CF_OLF : Data Protector Oracle Server integration and Data Protector SAP
R/3 integration-related options.
NOTE:
This option is not supported on EMC.
The default value is 0 . Possible values are 0 and 1 .
If an offline backup is performed using the Data Protector SAP R/3 integration, the option is ignored and the integration behaves as if the option was set to 1 .
Instant recovery
The instant recovery process depends on whether the control file and redo logs reside on the same disk array source volume as datafiles or not: l
If this option is set to
0
(default), during a ZDB session, Data Protector creates target volumes only for the source volumes containing Oracle datafiles. Target volumes for source volumes containing
Oracle control file and Oracle online redo logs are not created.
For Oracle proxy-copy or backup set ZDB and restore concepts when this option is set to
0
, see
Oracle backup set ZDB concepts , on page 22
and Oracle proxy-copy ZDB concepts , on page 26
.
For SAP R/3 backup and restore concept when this option is set to 0
, see Integration concepts , on page 18
.
l
If this option is set to
1
, Data Protector creates target volumes for all source volumes containing
Oracle datafiles, Oracle control file, and if Oracle integration is used, Oracle online redo logs.
Data Protector (10.00) Page 408 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
IMPORTANT:
If the ZDB_ORA_INCLUDE_CF_OLF option is set to 1 the control files and redo logs are overwritten during instant recovery.
Opening the database on the backup system
To successfully open the database on the backup system for other purposes than Data Protector, l
With Oracle proxy-copy ZDB method, set this option to
1
.
l
With Oracle backup set ZDB method, used with the Oracle integration, you can always open the database on the backup system.
Prerequisites
The prerequisites for this option to be set to 1 are: l
Data Protector Oracle Server integration: Oracle datafiles, Oracle control file, and Oracle online redo logs must be installed on a disk array.
l
Data Protector SAP R/3 integration: Oracle datafiles and Oracle control file must be installed on a disk array.
and Oracle backup set ZDB and restore concept when the ZDB_ORA_INCLUDE_CF_
OLF option is set to 1, on the next page
for Oracle backup and restore concepts when this option is set to
1
.
for SAP R/3 backup and restore concepts when this option is set to 1 .
Oracle proxy-copy ZDB and restore concepts when the ZDB_ORA_INCLUDE_CF_OLF option is set to 1
Data Protector (10.00) Page 409 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
Oracle backup set ZDB and restore concept when the ZDB_ORA_INCLUDE_CF_OLF option is set to
1
SAP R/3 backup and restore concept when the ZDB_ORA_INCLUDE_CF_OLF option is set to 1 with online backup, or in case of offline backup
Data Protector (10.00) Page 410 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix
ZDB_ORA_INCLUDE_SPF : Data Protector Oracle Server integration-related options.
The default value is 0 . Possible values are 0 and 1 .
The option is ignored and the integration behaves as if the option was set to 1 if offline backup is performed using the Data Protector SAP R/3 integration.
If this option is set to
0
, during ZDB sessions, Data Protector checks if Oracle Server datafiles and the
Oracle Server SPFILE are located on the same source volume. If the datafiles and the SPFILE are located on the same volume and instant recovery is enabled in the backup specification, the ZDB session fails. If this option is set to 1 , Data Protector skips the check. To enable instant recovery, leave this option set to the default value.
CAUTION:
If this option is set to 1 and the datafiles are located on the same volume as the SPFILE, the
SPFILE is overwritten during instant recovery, potentially resulting in a data loss.
ZDB_ORA_NO_CHECKCONF_IR:Data Protector Oracle Server integration-related option.
The default value is 0 . Possible values are 0 and 1 .
By default, the Oracle Server configuration is checked whether it is instant recovery-enabled or not
(whether the Oracle control file, the Oracle Server SPFILE, and the Oracle Server online redo logs are located on volumes of a different volume group than Oracle Server datafiles or not). For Oracle Server configuration check, the Data Protector command omniresolve
is used internally. On UNIX systems,
Data Protector (10.00) Page 411 of 413
Zero Downtime Backup Integration Guide
Chapter A: Appendix the omniresolve
file must have the setuid
bit set. When this option is set to
1
, the configuration check is omitted.
CAUTION:
Checking the Oracle Server configuration for instant recovery suitability is an essential step to ensure the instant recovery session does not result in a data loss. It is therefore not recommended to set this option to 1 unless you ensure that the Oracle Server is and remains configured appropriately for instant recovery.
OB2MARAWREAD_KB: This option sets the read block size for Oracle and SAP R/3 ZDB integrations on UNIX systems with Oracle tablespaces or datafiles installed on disk images and when using the proxy-copy method (when using DMA).
The default value is 64 kB. The specified value must be in the range between 1 kB and 1 MB.
The specified size is automatically adjusted to a size which is a multiple of the block size. The values above 256 kB could cause the DMA to fail.
ZDB_TAKE_CLUSRES_ONLINE: This option specifies how many times Data Protector tries to connect to Microsoft SQL Server in case the first connection fails. A reconnection is triggered every 30 seconds. This means that Data Protector waits up to ZDB_TAKE_CLUSRES_ONLINE x 30 seconds for the Microsoft SQL Server resources to start up.
SSEA_ATOMIC_SPLIT: Determines if the Data Protector HPE P9000 XP Agent should use the atomic split configuration of a disk array of the HPE P9000 XP Disk Array Family to ensure data consistency of replicas of the Oracle Server data in configurations where Automatic Storage
Management (ASM) is used.
Default:
0
(disabled). Possible:
0
|
1
.
Data Protector (10.00) Page 412 of 413
Send documentation feedback
If you have comments about this document, you can contact the documentation team by email. If an email client is configured on this system, click the link above and an email window opens with the following information in the subject line:
Feedback on Zero Downtime Backup Integration Guide (Data Protector 10.00)
Add your feedback to the email and click Send.
If no email client is available, copy the information above to a new message in a web mail client, and send your feedback to [email protected].
We appreciate your feedback!
Data Protector (10.00) Page 413 of 413
Download
Public link updated
The public link to your chat has been updated. You may view and manage your chat links after registration here.
Document public link
Advertisement
Key features
Zero downtime backups
Integration with various disk arrays
Standard and instant recovery options
Support for different database types
Backup scheduling and monitoring
Replica creation and management
Pre and post-backup scripting
Frequently asked questions
Data Protector is a software that enables zero downtime backup integration by creating replicas of the data that is being backed up. This allows you to perform backups without interrupting the operation of the application system.
Data Protector supports different types of disk arrays including HPE P6000 EVA Disk Array Family, HPE P9000 XP Disk Array Family.
The advantages of using Data Protector Oracle ZDB integration are that it reduces the performance degradation of the application system, the tablespaces are in backup mode only during a short period required to create a replica, and the load to the application system is significantly reduced.
