5. Instant Recovery configuration. Symantec NetBackup 7.5
Symantec NetBackup Snapshot Client 7.5 is a powerful backup and recovery solution that provides a variety of snapshot-based features for NetBackup. It supports clients on UNIX, Linux, and Windows platforms, on Fibre Channel networks (SANs) or traditional LANs. The Snapshot Client provides several features including Instant Recovery, off-host backup, FlashBackup, and RealTime.
Advertisement
Advertisement
Chapter
5
Instant Recovery configuration
This chapter includes the following topics:
■
About Instant Recovery capabilities
■
■
■
Giving full server privileges to the media server
■
■
Configuring a policy for Instant Recovery
■
About sizing the cache for Instant Recovery copy-on-write snapshots
■
■
Modifying the VxVM or FlashSnap resync options for point in time rollback
■
Instant Recovery for databases
■
About storage lifecycle policies for snapshots
About Instant Recovery capabilities
The Instant Recovery feature of Snapshot Client enables high-speed data retrieval from disk by means of the standard NetBackup user interface. Note the following features:
■ Supports NetBackup clients on Solaris, HP, AIX, Linux, and Windows. The master server can be on any supported operating system.
100 Instant Recovery configuration
Instant Recovery requirements
■ Uses snapshot technologies to create disk images.
■ Can create a snapshot and a backup to tape or disk, from one policy.
■ Enables random-access (non-sequential) restores of dispersed files from full backups.
■ Enables block-level restore and file promotion from VxFS_Checkpoint snapshots (UNIX) and file promotion from NAS_Snapshot. Also enables Fast
File Resync from VxVM and FlashSnap snapshots on Windows.
■ Enables rollback from the backups that were created using the following:
VxFS_Checkpoint, VxVM, FlashSnap, NAS_Snapshot, or disk array methods.
■ Can restore to a different path or host.
■ Provides the resource management by means of a rotation schedule.
■ Supports Oracle, Microsoft Exchange, DB2, SAP, and SQL-Server databases.
Instant Recovery requirements
Note the following requirements:
■ For snapshots using Storage Checkpoints, by means of NetBackup’s
VxFS_Checkpoint method, note that Solaris clients must have VxFS 3.4 or later (HP clients VxFS 3.5 or later, Linux and AIX clients VxFS 4.0 or later) with the Storage Checkpoint feature.
■ For VxVM snapshot volumes on UNIX, clients must have VxVM 3.2 or later with the FastResync feature. Linux and AIX clients require VxVM 4.0 or later.
Windows clients must have Storage Foundations for Windows version 3.1 or later.
■ For Instant Recovery with DB2, Oracle, Exchange, SAP, or SQL-Server databases, refer to the appropriate NetBackup database agent guide.
■ For replication hosts (using NetBackup’s VVR method), clients must have
VxVM 3.2 or later with the Veritas Volume Replicator feature. Linux and AIX clients require VxVM 4.0 or later.
Instant Recovery restrictions
Note the following restrictions:
■ For snapshots using Storage Checkpoints, Instant Recovery supports file systems with the Version 4 disk layout or later. Older disk layouts must be upgraded to Version 4 or later.
Instant Recovery configuration
Instant Recovery restrictions
101
■ No-data Storage Checkpoints (those containing file system metadata only) are not supported.
■ Instant Recovery snapshots must not be manually removed or renamed, otherwise the data cannot be restored.
■ Instant Recovery does not support the VxVM, FlashSnap, and VVR snapshot methods when used with VxVM volume sets.
■ On Linux, Instant Recovery is not supported by disk-array based snapshot methods.
■ For Instant Recovery backups of data that is configured on VxVM volumes on
Windows, the VxVM volume names must be 12 characters or fewer. Otherwise, the backup fails.
■ Any media server that is used in an Instant Recovery backup must have full server privileges.
See
“Giving full server privileges to the media server”
on page 102.
■ For Instant Recovery, Symantec strongly recommends that a primary volume be backed up by a single Instant Recovery policy. If the same volume is backed up by two or more Instant Recovery policies, conflicts between the policies may occur during snapshot rotation. Data loss could result if the policies are configured for snapshots only (if the policies do not back up the snapshots to separate storage devices).
Consider the following example: Two policies use the same snapshot device
(or VxFS storage checkpoint) to keep Instant Recovery snapshots of volume_1.
■
■
Instant Recovery policy_A creates a snapshot of volume_1 on the designated snapshot device or storage checkpoint.
When Instant Recovery policy_B runs, it removes the snapshot made by policy_A from the snapshot device or storage checkpoint. It then creates its own snapshot of volume_1 on the snapshot device or storage checkpoint.
The snapshot created by policy_A is gone.
Note: Even if each policy has its own separate snapshot devices, conflicts can occur when you browse for restore. Among the available snapshots, it may be difficult to identify the correct snapshot to be restored. It is therefore best to configure only one policy to protect a given volume when using the Instant
Recovery feature of NetBackup.
102 Instant Recovery configuration
Giving full server privileges to the media server
Giving full server privileges to the media server
A media server used in an Instant Recovery backup must have full server privileges.
If it does not have server privileges, the snapshots created by each backup do not properly expire in the NetBackup catalog.
To give full server privileges to the media server
1 In the NetBackup Administration Console, click Host Properties > Master
Server > double click master server > Master Server Properties > Servers.
2 Make sure that the media server is listed under Additional servers, not under
Media servers.
Note: on UNIX, this procedure places a SERVER = host entry in the bp.conf
file for each host that is listed under Additional servers. In the bp.conf
file, the media server must not be designated by a MEDIA_SERVER = host entry.
About Instant Recovery
Standard NetBackup can use disks for backup and restore. The Instant Recovery feature of Snapshot Client extends that capability by exploiting the speed of
Instant Recovery configuration
About Instant Recovery
103
NetBackup master server with Instant
Recovery policy
NetBackup client snapshots. A snapshot is created with minimal effect on client access to data or on the speed of client transactions. If snapshots are made frequently, you can restore an accidentally deleted file in a matter of seconds.
shows an example.
Figure 5-1 Instant Recovery from snapshot on disk
Client’s working data
Snapshots on same or different disk snapshot A snapshot B snapshot C
Time of creation
Dec 9 07:30:16
2004
Dec 9 08:30:11
2004
Dec 9 09:30:15
2004 restore
The on-disk snapshots become point-in-time versions of file systems or volumes, to be kept or discarded as needed. The data can be restored from local disk; no need to mount a tape or other remote storage.
NetBackup Instant Recovery creates snapshot A of the client data on disk. One hour later, as scheduled, NetBackup creates snapshot B, also on disk, followed one hour later with snapshot C. By using the appropriate snapshot, you can restore data directly from disk.
Note: NetBackup Instant Recovery retains the snapshot. The snapshot can be used for restore even if the client has been restarted.
The following sections provide some background information.
About snapshot and backup for Instant Recovery
An Instant Recovery backup creates a snapshot on disk and optionally backs up the client’s data to a storage device. The location of the snapshot depends on the type of snapshot method that is configured in the policy.
About NetBackup catalog maintenance
For Instant Recovery backups, NetBackup automatically updates the catalog to keep it correlated with the snapshots on the client. If not kept up-to-date, the
104 Instant Recovery configuration
About Instant Recovery catalog might refer to snapshots that no longer exist on the client, due to user activity (snapshot replacement or deletion).
NetBackup includes a maintenance program ( bppficorr ) that runs after backups and restores. It can also be run manually to update the catalog if an Instant
Recovery snapshot is accidentally deleted or renamed.
For man page information on the bppficorr command, refer to the NetBackup
Commands guide.
About snapshot management
Because snapshots require disk space, they cannot be retained forever. To balance the space consumed against the convenience of having multiple snapshots available for instant recovery, you can specify how many snapshots to retain. For many disk array snapshot methods, you can also specify the particular devices on which to retain the snapshots, and the order in which they are used.
About snapshot rotation
NetBackup Snapshot Client implements a rotation scheme for managing snapshot resources for some snapshot methods.
In the following bulleted example, the system is configured to have three snapshots available for instant restore ("undo") at any time:
■ The state of the client’s backup selection list is captured in the first Instant
Recovery snapshot, and the snapshot is retained. In other words, it is not deleted, even if the backup image is also copied to a storage unit.
■ When the backup policy runs again, recent changes in the client’s backup selection list are captured in a second snapshot. This snapshot is created on a second device that was allocated for this purpose. The first snapshot is retained, in case a restore is needed from that data.
■ When the backup policy runs a third time, a third snapshot is created on the third allocated device. At this point three snapshots representing three different states of the client data are available for restore.
■ When the policy runs a fourth time, no additional snapshot device is available: one of the existing devices must be reused. The system "rotates" the devices, overwriting the first snapshot with the fourth (most recent) one. Although the earliest snapshot is now gone, the three most recent snapshots are available for restore.
Figure 5-2
Instant Recovery configuration
About Instant Recovery
105
Snapshot rotation, for multiple restore (undo) levels
Client’s backup selections
Oldest snapshot:
12:00 noon
Most recent snapshot:
4:00 PM
Older snapshot: 2:00 PM
In this figure, the next Instant Recovery backup overwrites the snapshot that was made at 12:00 noon.
Means of controlling snapshots
Depending on the snapshot method you have selected for the policy, one of two means of managing snapshots is available: the Snapshot Resources pane or the
Maximum Snapshots (Instant Recovery only) parameter.
Snapshot Resources pane
The Snapshot Resources pane (on the policy’s Snapshot Options dialog box) is available only for certain disk array snapshot methods on UNIX clients.
The Snapshot Resources pane provides the most control over Instant Recovery snapshots. You can specify which devices to use for the snapshots, and the order in which the devices are used. Click the Add option to display the following:
Figure 5-3 Example of Add Snapshot Resource dialog
106 Instant Recovery configuration
About Instant Recovery
■ The serial number of the array is specified in the Array Serial # field. Contact your array administrator to obtain the disk array serial numbers and designators (unique IDs) for the array.
The unique ID snapshot resource or source LUN containing the primary data is specified in the Source Device.
The maximum number of snapshots to retain is determined by the number of configured devices in the Snapshot Device(s) field. For example, if you enter two devices, only two snapshots can be retained. The above example specifies three devices (0122;0123;0124), so three snapshots can be retained. When the maximum is reached, the fourth snapshot overwrites the first one.
■ The particular devices to use for the snapshots are those named in the Snapshot
Device(s) field.
■ The order in which the devices are listed in the Snapshot Device(s) field determines their order of use. Device 0122 is used for the first snapshot, 0123 for the second, and 0124 for the third.
Preconfiguration of the snapshot devices may be required.
See the appropriate topic for your disk array and snapshot method.
Maximum Snapshots parameter
The Maximum Snapshots parameter on the Snapshot Options dialog box sets the maximum number of Instant Recovery snapshots to be retained at one time.
This parameter is available only for certain snapshot methods. Unlike the Snapshot
Resources pane, it cannot specify which devices to use or the order of their use.
(The Maximum Snapshots parameter and the Snapshot Resources pane are mutually exclusive.)
When the maximum is reached, the next snapshot causes the oldest to be deleted.
Note: For Windows clients using the VSS method on disk arrays that are configured for clones or mirrors: you must synchronize the clones or mirrors with their source before you run the backup.
For Instant Recovery backups, it is good practice to set the backup retention level to infinite. A retention period that is too short can interfere with maintaining a maximum number of snapshots for restore.
Instant Recovery configuration
Configuring a policy for Instant Recovery
107
Configuring a policy for Instant Recovery
This section explains how to configure a policy for the backups that are enabled for Instant Recovery. As an alternative to this procedure, you can use the Policy
Configuration Wizard.
To configure a policy for Instant Recovery
1 Start the NetBackup Administration Console as follows:
On UNIX, enter:
/usr/openv/netbackup/bin/jnbSA &
On Windows, click Start > Programs > Symantec NetBackup > NetBackup
Administration Console.
2 Click Policies. In the All Policies pane, open a policy or create a new one.
3 For the policy type, select Standard, MS-Windows, FlashBackup-Windows, or the database agent type appropriate for the client(s).
4 Select a storage unit (disk or tape).
If you select Snapshots only on the Schedules tab, the storage unit is not used. NetBackup creates a snapshot only.
5 Select Perform snapshot backups.
6 Select Retain snapshots for instant recovery.
NetBackup retains the snapshot so that Instant Recovery can be performed from the snapshot. A normal backup to storage is also performed, if you do not select Snapshots only on the Schedules tab.
7 To save the settings, click Apply.
8 Click the Snapshot Client Options option to select the snapshot method.
For a new policy, you can skip this step to let NetBackup select the method
(auto is the default method).
■ Select a snapshot method from the pull-down list. For creating an Instant
Recovery snapshot, the available methods are: auto
(UNIX or Windows)
NetBackup selects the snapshot method. If auto is the method, only one snapshot can be retained at a time.
108 Instant Recovery configuration
Configuring a policy for Instant Recovery
FlashSnap
(UNIX or Windows)
NAS_Snapshot
(UNIX or Windows)
OST_FIM
VSS
(Windows 2003 only)
VxFS_Checkpoint
(UNIX only)
VxVM
(UNIX or Windows)
Uses the VxVM FlashSnap feature and VxVM
FastResync to create the snapshot. VxVM mirrors must be configured:
See
on page 112.
This method can also use instant snapshots:
See
“About VxVM instant snapshots”
on page 168.
For Windows clients, FlashSnap can use the Fast
File Resync feature of Storage Foundation for
Windows 4.1.
Uses the NDMP V4 snapshot extension to create the snapshot on the NAS-attached disk.
See
“Setting up a policy for NAS snapshots”
on page 156.
The name of the snapshot method that is selected in a policy configured for snapshot replication using Replication Director. The name represents
OpenStorage Frozen Image Method.
Refer to the Symantec NetBackup™Replication
Director Solutions Guide for more details.
VSS selects the actual snapshot method depending on which snapshot provider is configured on the client. For example, if the client data is stored on an EMC CLARiiON disk array, and the array administrator has fully configured the array and its snapshot capability: VSS selects the appropriate
EMC CLARiiON snapshot method for the backup.
Uses VxFS Storage Checkpoint to create the snapshot. A new Storage Checkpoint is created whenever a backup using the VxFS_Checkpoint method is executed.
Uses VxVM FastResync to create the snapshot.
VxVM mirrors must be configured:
See
on page 112.
The VxVM method can also use instant snapshots:
See
“About VxVM instant snapshots”
on page 168.
For Windows clients, VxVM can use the Fast File
Resync feature of Storage Foundation for Windows
4.1.
Instant Recovery configuration
Configuring a policy for Instant Recovery
109
VVR
(UNIX)
The disk array snapshot methods.
Creates a snapshot of a VxVM volume on a Veritas
Volume Replication host.
See the appropriate topic for your disk array and snapshot method.
■ Change parameter values for the method, if needed.
■ When you are finished, click OK.
9 To configure a schedule, use the Schedule tab.
■ For a snapshot only, select Snapshots only.
If Snapshots only is selected, the snapshot is not backed up to tape or other storage. NetBackup creates a snapshot on disk only. This option is required for the NAS_Snapshot method. Note that you must deselect
Snapshots only if you want to deselect Retain snapshots for instant
recovery on the policy Attribute tab.
If the snapshot uses VxFS_Checkpoint or is a VxVM space-optimized snapshot, it is created on the same device as the device containing the original data. In that case, you may want to create another policy to back up the data to a separate device.
If Snapshots and copy snapshots to a storage unit is selected, NetBackup creates (and retains) a snapshot and backs up the client’s data to the storage unit that is specified in the policy.
■ You can select the retention period for snapshots under Retention.
■ Make other schedule selections as desired, and click OK.
10 To enter the files and folders to be backed up, use the Backup Selections tab.
■ When backing up Oracle database clients, refer to the NetBackup for Oracle
System Administrator’s Guide for instructions.
■ Snapshot Client policies do not support the ALL_LOCAL_DRIVES entry in the policy’s Backup Selections list, except for the policies that are configured with the VMware method.
■ If you use the Policy Configuration Wizard:
See
“About the Policy Configuration Wizard”
on page 81.
11 To specify clients to be backed up by this policy, use the Clients tab.
12 On the policy Attributes tab: if you click Apply or OK, a validation process checks the policy and reports any errors. If you click Close, no validation is performed.
110 Instant Recovery configuration
About sizing the cache for Instant Recovery copy-on-write snapshots
About sizing the cache for Instant Recovery copy-on-write snapshots
A copy-on-write snapshot requires cache space for storing the changes that occur on the source device during the life of the snapshot. While the snapshot is active, any blocks that are about to be changed by user activity are copied to the cache.
Blocks that do not change on the source are not copied. Compared to a fully-allocated snapshot (clone or mirror), a copy-on-write snapshot may consume relatively little disk space and can be executed very quickly.
As a rule, the appropriate size for the cache depends on the amount of user activity that occurs during the life of the snapshot. The more changes in the source data, or the longer the life of the snapshot, the more blocks that are likely to be changed.
As a result, more data must be stored in the cache.
The size of the file system or raw partition does not determine cache size. If little change activity occurs on the source during the life of the snapshot, little cache space is required, even for a large file system.
Note: If the cache runs out of space, the snapshot may fail.
For more detail on the copy-on-write process, refer to How copy-on-write works in the NetBackup Snapshot Client Administrator's Guide.
Cache size during restore
The size of the cache may need to increase when you restore large amounts of data from an Instant Recovery snapshot. For instance, if you restore an entire raw partition to a source device that has an active Instant Recovery snapshot, writing the restore data causes the active snapshot to cache all blocks from the raw partition that resides on the source device. A cache that was large enough when the snapshot was first initiated may no longer be large enough. The large restore multiplies the write-to-cache activity, increasing the space that is needed for cache.
Use the following copy-on-write snapshot methods with Instant Recovery only when small amounts of data (individual files) are likely to be restored:
■ VxVM space optimized instant snapshots
■ VxFS_Checkpoint
■ The following copy-on-write disk array methods if the snapshot is not fully allocated:
■ EMC_CLARiiON_SnapView_Snapshot
Instant Recovery configuration
About sizing the cache for Instant Recovery copy-on-write snapshots
111
■ EMC_TimeFinder_Snap
■ HP_EVA_Snapshot
■ HP_EVA_Vsnap
■ Hitachi_CopyOnWrite
Setting an adequate size for the snapshot cache
The cache must be sized appropriately in either of the following cases:
■ If a lot of write activity is expected in the source data during the life of the copy-on-write snapshot.
■ If you expect to restore large amounts of data from an Instant Recovery snapshot.
To set an adequate size for the snapshot cache
◆ Use the following formula:
For raw partitions:
For file systems:
Cache size = volume size * the number of retained snapshots
Cache size = (consumed space * the number of retained snapshots)
+ approximately 2% to 5% of the consumed space in the file system
Note:
■
■
Consumed space means the total size of the data allocated to files in the file system, not the total size of the mounted volume.
The additional space (2% to 5%) represents a safety measure and may not be required.
Large restores from an Instant Recovery snapshot
For restore of an entire raw partition or large file system, Symantec recommends a fully allocated snapshot method.
Examples are the following:
■ VxVM as a snapshot mirror (not space-optimized)
■ EMC_CLARiiON_Snapview_Clone
■ EMC_TimeFinder_Clone
■ EMC_TimeFinder_Mirror
■ HP_EVA_Snapclone
112 Instant Recovery configuration
About configuring VxVM
■ Hitachi_ShadowImage
■ IBM_DiskStorage_FlashCopy
■ IBM_StorageManager_FlashCopy
■ Any copy-on-write disk array method that is configured with a fully-allocated snapshot.
About configuring VxVM
For Instant Recovery backups of the data that is configured on VxVM volumes on
Windows, the VxVM volume names must be 12 characters or fewer. Otherwise, the backup fails.
Before you run an Instant Recovery policy to back up VxVM volumes, one or more mirrors must be created. The primary volumes must be enabled for FastResync.
Note that on Windows, FastResync is enabled by default.
Creating a snapshot mirror
You can create a snapshot mirror using VxVM commands. As an alternative, you can use the Policy Configuration Wizard in the NetBackup Administration Console.
To create a snapshot mirror on Windows
1 For a volume that is associated with a drive letter, enter: vxassist snapstart X: where X is the drive letter. This command creates a snapshot mirror of the designated drive.
2 For a volume that is not associated with a drive letter, enter: vxdg -g disk_group dginfo
This command shows information for the specified disk group, including the names of the volumes that are configured for that group.
■ Create the snapshot by entering the following: vxassist snapstart \Device\HarddiskDmVolumes\disk_group\Volume_name
This command creates a snapshot mirror of the designated Windows volume.
Instant Recovery configuration
About configuring VxVM
113
To create a snapshot mirror on UNIX
1 Add dco (data change object) logs to the primary volume:
/usr/sbin/vxassist -g disk_group addlog volume_name logtype=dco
2 Enable FastResync on the volume:
/usr/sbin/vxvol -g disk_group set fmr=on volume_name
3 Prepare a new mirror for the Instant Recovery snapshot:
■ Create the mirror:
/usr/sbin/vxassist -g disk_group snapstart primary_volume
Wait until the mirror is synchronized (status SNAPDONE, or State field reads Ready in the volume’s properties display).
■ To verify that the mirror is synchronized, enter:
/usr/sbin/vxprint -g disk_group -q -t -e 'assoc="primary_volume"'
About creating Instant Snapshots
The following procedures are required for Instant Recovery backups when using the full-sized or space-optimized instant snapshot options in VxVM 4.0. For Instant
Recovery backups, VxVM 4.0 or later instant snapshots are supported by Snapshot
Client’s FlashSnap, VxVM, and VVR methods.
Creating space-optimized snapshots
A cache object called NBU_CACHE must be created in the disk group containing the volume to be backed up. NetBackup recognizes the cache object and uses it to create a space-optimized snapshot.
You can use the Policy Configuration Wizard for these steps when you create a new policy.
To create space-optimized snapshots
1 Create the parent volume:
/usr/sbin/vxassist -g disk_group make volume size layout=layout logtype=dco dcoversion=20 [drl=no|sequential|yes]
[ndcomirror=number] fastresync=on
Where:
114 Instant Recovery configuration
About configuring VxVM
■ Brackets [ ] indicate optional items.
■ make volume specifies the name of the volume snapshot.
2 Create the cache object:
/usr/sbin/vxassist -g disk_group make cache_volume size layout=layout init=active
3 Label the cache object:
/usr/sbin/vxmake -g disk_group cache NBU_CACHE cachevolname=cache_volume
4 Enable the cache object:
/usr/sbin/vxcache -g disk_group start NBU_CACHE
5 Take the initial snapshot:
/usr/sbin/vxsnap -g disk_group make source=volume/newvol
=SNAP_vol1_NBU/cache=NBU_CACHE
Creating full-sized snapshots
Unlike the space-optimized snapshot, NetBackup cannot create VxVM full-sized instant snapshots: you must create them before running the backup, as explained in the following procedure. You must create one full-sized instant snapshot for each backup you want to run.
As an alternative, you can use the Policy Configuration Wizard to perform these steps when creating a new policy.
To create full-sized snapshots
1 Enter the following to create the parent volume:
/usr/sbin/vxassist -g disk_group make volume length layout=layout logtype=dco dcoversion=20 [drl=no|sequential|yes]
[ndcomirror=number] fastresync=on
Where:
■ Brackets [ ] indicate optional items.
Instant Recovery configuration
About configuring VxVM
115
■ make volume specifies the name of the volume snapshot.
2 Create a volume for a full-sized instant snapshot:
■ Determine the required size for the snapshot volume:
# LEN='vxprint -g disk_group -F%len volume'
■ Find the name of the DCO volume:
# DCOVOL='vxprint -g disk_group -F%dconame volume'
■ Discover the DCO volume's region size (in blocks):
# RSZ='vxprint -g disk_group -F%regionsz $DCOVOL'
■ Create a volume that is named volumename_NBU, of the required size and redundancy.
The volume name must end with
_NBU
. In the following example, the volume is named
SNAP_vol1_NBU
.
vxassist -g disk_group make SNAP_vol1_NBU $LEN layout=mirror nmirror=number logtype=dco drl=no dcoversion=20 ndcomirror=number regionsz=$RSZ init=none
[storage attributes ...]
The number for nmirror should equal the number for ndcomirror
.
Note: For Linux, the init value should be init=active instead of init=none.
For Solaris 10 with Storage Foundation 5.1, the init value should be init=active instead of init=none.
■ Create the mirror: vxsnap -g disk_group make source=volume/snapvol=SNAP_vol1_NBU/syncing=on
3 Set the Maximum Snapshots (Instant Recovery only) value on the NetBackup
Snapshot Client Options dialog.
Using the VxVM 3.5 graphical user interface to configure VxVM mirrors
The following are the steps for using the VxVM 3.5 graphical user interface to configure VxVM mirrors for Instant Recovery backups.
116 Instant Recovery configuration
About configuring VxVM
To use the VxVM 3.5 graphical user interface to configure VxVM mirrors
1 Make sure that FastResync is enabled on the primary VxVM volume.
■ In the VEA console, right-click the volume and click Properties from the pop-up menu.
■ The FastResync field states whether or not FastResync is enabled.
■ Click Cancel.
■ If FastResync is disabled, right-click the volume again and select Fast
Resync > Add from the pop-up menu to enable it.
2 Prepare a new mirror for Instant Recovery backup.
■ Create a new mirror by right-clicking the volume and selecting Snap >
Snap Start.
■ Make sure FastResync is enabled on the mirror. Click OK to create the mirror and start full synchronization.
Instant Recovery configuration
About configuring VxVM
117
3 On the Mirrors tab, ensure that synchronization has completed as indicated by Snap Ready in the Status field.
118 Instant Recovery configuration
Modifying the VxVM or FlashSnap resync options for point in time rollback
Modifying the VxVM or FlashSnap resync options for point in time rollback
Several options can be set for Instant Recovery point in time rollback when multiple volumes are involved in the rollback. For better performance, you can modify these options to suit the circumstances of the rollback.
The following options can be modified for rollback of a VxVM or FlashSnap snapshot:
■ Maximum number of volumes to resynchronize concurrently
■ Sync region size in MB (UNIX only)
■ Sync I/O delay in milliseconds (UNIX only)
By default, these resync options have the same values for restore as they had for the backup. For the defaults, see the following section:
See
“Configuration parameters for Snapshot Client”
on page 69.
To modify the VxVM or FlashSnap resync options for point in time rollback
1 Create the following file:
/usr/openv/netbackup/SYNC_PARAMS
2 In the file, enter the numeric values for the options, on one line. The numbers apply to the options in the bulleted list above, in that order.
For example:
6 3 1000
This example resets the options as follows:
■ Maximum number of volumes to resynchronize concurrently = 6
■ Sync region size in MB (UNIX only) = 3
■ Sync I/O delay in milliseconds (UNIX only) = 1000
Instant Recovery for databases
To configure an Instant Recovery policy for database clients, refer to the appropriate NetBackup database agent guide.
Instant Recovery configuration
About storage lifecycle policies for snapshots
119
About storage lifecycle policies for snapshots
A storage lifecycle policy is a storage plan for a set of backups. NetBackup uses the lifecycle policy to determine where to store additional copies of the backup images and how long to retain those copies. In general, short-term copies can be kept on disk (for quick restore) and long-term copies can be kept on tape or other storage.
NetBackup can manage snapshot-based backups for Instant Recovery through storage lifecycle policies.. The Instant Recovery feature makes snapshots available for quick data recovery from disk. Lifecycle policies support a lifecycle storage plan for the storage unit copies that are made during an Instant Recovery backup.
Note that lifecycle policies with Instant Recovery snapshots are especially useful with the NetBackup RealTime product.
Lifecycle policy example with RealTime
shows an example hardware configuration for a lifecycle policy. The lifecycle policy specifies three destinations and three retention periods for managing RealTime snapshot-based backups.
Figure 5-4
NetBackup master server
Lifecycle policy example with RealTime
LAN / WAN
NetBackup client
Media server and alternate client
Tape library
3
Fibre channel SAN
Real Time storage
1
Real Time storage
2
Diskarray
Destination 1: Destination 2: Destination 3:
Net Backup creates a consistent point-in-time state in Real Time storage.
Point-in-time date is retained in the Real Time storage for
2 weeks, available for Instant
Recovery.
The Net Backup media server creates a backup image from a Real Time snapshot and copies the image to the disk array on the SAN.
The Net Backup media server duplicates the disk image and sends the duplication to tape.
For data that is more than 6 months old, recovery can be made from tape.
For data that is more than 2 weeks old, recovery can be made from the array.
120 Instant Recovery configuration
About storage lifecycle policies for snapshots
Configuring a storage lifecycle policy to manage snapshot-based backups for Instant Recovery
This section desceribes how to create a storage lifecycle policy to manage snapshot-based backups for Instant Recovery. The procedure focuses on snapshot-related details only.
Full procedures are available for creating storage lifecycle policies:
See the NetBackup Administrator’s Guide, Volume I.
To configure a storage lifecycle policy to manage snapshot-based backups for
Instant Recovery
1 Create a lifecycle policy with two or more storage destinations.
Use the Storage > Storage Lifecycle Policies node of the NetBackup
Administration Console. Click Actions > New > Storage Lifecycle Policies.
Click Add.
Here are suggestions for a lifecycle policy for RealTime Protection and multiple storage destinations:
■ For snapshots, select Snapshot on the New Storage Destination dialog.
You can specify a retention period appropriate for snapshots (such as two weeks). Click OK.
Note that storage units are not used for snapshots. Storage units are used only for backup (and duplication) images.
■ For backup copies to disk, select Backup on the New Storage Destination dialog. Specify a disk storage unit and a longer retention period (such as six months). Click OK.
■ For backup copies to tape, select Duplication on the New Storage
Destination dialog. Specify a tape storage unit and a longer retention period (such as five years). Click OK and finish creating the lifecycle policy.
2 Create a policy for snapshots. (Use the Policies node of the Administration
Console.)
On the policy Attributes tab:
■ You can specify the lifecycle policy in the Policy storage unit / lifecycle
policy field. You can later change the lifecycle policy in the schedule, as explained later in this procedure.
■ Select Perform snapshot backups.
■ For RealTime Protection with lifecycles, Symantec recommends selecting
Retain snapshots for Instant Recovery. This option is required if you want NetBackup to keep the snapshot after the backup occurs.
Instant Recovery configuration
About storage lifecycle policies for snapshots
121
■ Click Snapshot Client Options. To use RealTime Protection, select the
CDP snapshot method for Solaris or VSS for Windows. (Other methods can be used if you are not backing up a client that uses NetBackup
RealTime Protection.)
■ On the Snapshot Options dialog box, the Maximum Snapshots (Instant
Recovery only) parameter sets the maximum number of snapshots to be retained at one time. When the maximum is reached and another snapshot is created, the oldest snapshot is deleted.
Note that if you also set a snapshot retention period of less than infinity in the lifecycle policy, the snapshot is expired when either of these settings takes effect (whichever happens first). For example, if the Maximum
Snapshots value is exceeded before the snapshot retention period that is specified in the lifecycle policy, the snapshot is deleted.
The same is true for the Snapshot Resources pane on the Snapshot
Options dialog box. If the snapshot method requires snapshot resources, the maximum number of snapshots is determined by the number of devices that are specified in the Snapshot Device(s) field. For example, if two devices are specified, only two snapshots can be retained at a time. Either the Snapshot Device(s) field or the snapshot retention period in the lifecycle policy can determine the retention period.
3 Create a schedule for the policy.
The figure in this topic shows that a lifecycle configuration with three storage destinations is an intelligent means of managing CDP snapshot-based backups:
See
“Lifecycle policy example with RealTime”
on page 119.
You can create a single schedule for backups and let the lifecycle policy govern their destinations and retention periods, as follows:
■ Under Destination, if you selected Retain snapshots for Instant Recovery on the policy Attributes tab, make sure that Snapshots and copy
snapshots to a storage unit is selected on the schedule (not Snapshots
only).
Important: if you select Snapshots only on the schedule, a lifecycle policy cannot be used.
■ In the Override policy storage selection field, select the lifecycle policy that you created in
■ Under Schedule type, set an appropriate frequency, such as 1 day.
When the Snapshot Client policy executes this schedule, the lifecycle policy named in the Override policy storage selection field creates images on the destinations that are named in the lifecycle policy. The lifecycle policy also
122 Instant Recovery configuration
About storage lifecycle policies for snapshots sets the retention periods for the images it creates. In this example, the retention is six months for backups to disk and five years for tape.
4 To make full use of RealTime Protection with Instant Recovery, create another schedule to make CDP snapshots more frequently. This schedule will not be managed by lifecycle policies.
Specify the following:
■ Under Schedule type, compare to the schedule that is created at
a higher frequency (such as 1 hour).
■ Under Destination, select Snapshots only.
■ Do not specify a lifecycle policy for this schedule. Likewise, do not specify a lifecycle policy under the Policy storage unit / lifecycle policy field of the policy Attributes tab. Lifecycle policies do not manage snapshots without backups.
■ For Retention, specify the same period that you set for the Snapshot destination on the New Storage Destination dialog (see
the time period is two weeks.
With this configuration, the snapshot-only schedule creates a CDP snapshot every hour. The other schedule (which specifies the lifecycle policy under the
Override policy storage selection field) creates a snapshot and backup to disk and tape once per day. This configuration provides an inexpensive means of continuous data protection.
More detailed information on using NetBackup RealTime Protection is available.
See
on page 128.
Storage lifecycle polices and Snapshot Client troubleshooting
The section includes information about various error messages related to SLPs and snapshots.
If you configure a snapshot method for a policy, and the schedule specifies a lifecycle policy, the life cycle policy must include a snapshot destination.
Otherwise, an error such as the following appears in the NetBackup Problems report: snapshot backup: tashina11_1204305543 cannot be used with a lifecycle policy NoSnapshot that does not include a snapshot destination.
Error 156 can be a result of different problems, listed below are some of them:
Instant Recovery configuration
About storage lifecycle policies for snapshots
123
VxVM failing to get the version of the disk group, run appropriate VxVM command outside of NetBackup to see whether you can get the version information for the disk group in use.
Bpfis log
10:43:58.436 [28336] <32> onlfi_fim_dev_info: FTL - VfMS error 11; see following messages:10:43:58.437 [28336] <32> onlfi_fim_dev_info:
FTL - Fatal method error was reported
10:43:58.437 [28336] <32> onlfi_fim_dev_info:
FTL - vfm_freeze: method: vxvm, type: FIM, function: vxvm_freeze
10:43:58.437 [28336] <32> onlfi_fim_dev_info: FTL - VfMS method error 10; see following message:
10:43:58.437 [28336] <32> onlfi_fim_dev_info:
FTL - vxvm__get_dgversion: Cannot get version for disk group: dgdb001
10:43:58.437 [28336] <4> onlfi_thaw:
INF - Thawing /ora/db001/data001 using snapshot method vxvm.
10:43:58.448 [28336] <4> onlfi_thaw: INF - do_thaw return value: 0
10:43:58.454 [28336] <16> bpfis:
FTL - snapshot preparation failed, status 156
The device that is to backup by this process is being used by another process.
Check whether any other process is holding the same device.
Bpfis log
00:26:19.025 [2826] <2> onlfi_vfms_logf: INF - lock pid(2902) != pid(2826):
/usr/openv/netbackup/online_util/db_cntl/__LOCKFILE_EMC:
SYMMETRIX:970960001000
00:26:19.025 [2826] <2> onlfi_vfms_logf: INF - TimeFinder_rebuild:
Cannot get lock on device: /dev/rdsk/c3t5006048C4A85A400d1s2 ......
00:26:19.025 [2826] <32> rebuild_fim_list: FTL - TimeFinder_rebuild:
Cannot get lock on device: /dev/rdsk/c3t5006048C4A85A400d1s2
00:26:19.025 [2826] <32> splthost_rebuild: FTL - rebuild_fim_list() failed
00:26:19.037 [2826] <4> bpfis Exit:
INF - EXIT STATUS 156: snapshot error encountered
Policy validation fails for valid backup selection. If there filer’s volume is mounted on a windows client, run NetBackup client service on the client and the alternate client with a valid credentials to access CIFS share, and check that the filers are up, and the volume is seen mounted on the windows client.
Bpfis log
11:49:40.727 [15240.13716] <16> bpfis main: FTL - process_fs_list() failed, status 71
124 Instant Recovery configuration
About storage lifecycle policies for snapshots
11:49:40.727 [15240.13716] <2> ol_cleanup:
INF - removing
C:\Program Files\Veritas\NetBackup\temp\unknown+15240+1.std_filelist
11:49:40.727 [15240.13716] <4> bpfis Exit:
INF - EXIT STATUS 71: none of the files in the file list exist
11:49:40.743 [15240.13716] <2> stop_keep_alive_thread:
INF - Stop keep_alive thread
11:49:40.743 [15240.13716] <2> bpfis Exit: INF - Close of stdout
For windows client, live browse from the snapshot fails with the following error message. Make sure that NetBackup client service on the client and the alternate client is running with a valid credential to access CIFS share
ERROR: permissions denied by client during rcmd.
Snapshot backup for windows client fail with status 55. Make sure that NetBackup client service on the client and the alternate client is running with a valid credential to access CIFS share.
Bpfis log
10:46:00.131 [2612.7880] <4> bpcd_request_mount: get volume guid from <frag-id> failed with 55
10:46:00.131 [2612.7880] <4> bpcd_request_mount: return mntdev NO_MOUNT_DEVICE
10:46:00.131 [2612.7880] <16> bpcd_get_fileinfo: bpcd_request_mount failed with error[55]
10:46:00.147 [2612.7880] <2> bpcd_get_fileinfo: sent status 55 to bpdbm
Live browse or ‘backup from snapshot’ operation for windows client fail with error
43, status 156. Enable create_ucode & convert_ucode on primary volume.
Bpfis log
04:43:44.656 [3040.3900] <2> onlfi_vfms_logf:
INF - snapshot services:ostfi:Wed Aug 24 2011
04:43:44.640000 <Thread id - 3900> Failed to import snapshot [*************]
04:43:44.718 [3040.3900] <2> onlfi_vfms_logf:
INF - snapshot services: ostfi:Wed Aug 24 2011
04:43:44.718000 <Thread id - 3900> OST Library call failed with message
(STS API sts_create_export failed with error code : 2060022)
04:43:44.718 [3040.3900] <2> onlfi_vfms_logf: INF - snapshot services: ostfi:Wed Aug 24 2011
04:43:44.718000 <Thread id - 3900> COSTPlugin::importTreeNode -
Could not import device[************]
Instant Recovery configuration
About storage lifecycle policies for snapshots
125
NBUAdapter log
0 RESTORE :2104 111 0 115851 2011/08/24 04:58:51
Volume name = f3070-238-15:/NetBackup_1314174973_mirror, dest snap f270-247-156_test6
1 RESTORE :2104 44 0 115851 2011/08/24 04:58:51 share name:NBU_Share_NetBackup_1314174973_mirror_ f270-247-156_test6_2011_08_24_04_36_14
0 RESTORE :2104 6 0 115851 2011/08/24 04:58:51 add_cifs_export :
Failed to get response. Error 22 : Directory
"/vol/NetBackup_1314174973_mirror/.snapshot/f270-247-156_test6" does not exist.
1 STRWIDE :2104 1 0 115851 LFB:
Need 111 chars to store wide copy of UTF8
'Directory "/vol/NetBackup_1314174973_mirror/.snapshot/f270-247-156_test6" does not exist. ‘
126 Instant Recovery configuration
About storage lifecycle policies for snapshots
Advertisement
Key features
- Instant Recovery
- Off-host backup
- FlashBackup
- RealTime
- Block level incremental backup (BLIB)
- Support for Cluster Volume Manager Environments (CVM)
- Support for multiple disk arrays
- NAS snapshot