SSR212MA SAN SW Release Notes

Intel® Storage System
SSR212MA
Software Release Notes
Revision 1.4
Jan, 2007
Storage Systems Technical Marketing
Revision History
Intel® Storage System SSR212MA
Revision History
ii
Date
Revision
Number
December, 2005
1.0
Modifications
1st Release copy.
January, 2006
1.1
Added 6 additional issues: 2.1, 3.8, 3.9, 3.10, 11.12, 12.4.
June, 2006
1.2
Added 6.3 SP1 updates & issues, noted by “(SP1)” in the issue title.
August, 2006
1.3
Added 6.5 updates & issues
January, 2007
1.4
Added 6.6 updates & issues
Revision 1.4
Intel® Storage System SSR212MA
Disclaimers
Disclaimers
THE INFORMATION IN THIS DOCUMENT IS PROVIDED "AS IS" WITH NO WARRANTIES WHATSOEVER, INCLUDING ANY
WARRANTY OF MERCHANTABILITY, FITNESS FOR ANY PARTICULAR PURPOSE, OR ANY WARRANTY OTHERWISE
ARISING OUT OF ANY PROPOSAL, SPECIFICATION, OR SAMPLE.
®
Information in this document is provided in connection with Intel products. No license, express or implied, by estoppel or otherwise,
to any intellectual property rights is granted by this document. Except as provided in Intel's Terms and Conditions of Sale for such
products, Intel assumes no liability whatsoever, and Intel disclaims any express or implied warranty, relating to sale and/or use of
Intel products including liability or warranties relating to fitness for a particular purpose, merchantability, or infringement of any
patent, copyright or other intellectual property right. Intel products are not intended for use in medical, life saving, or life sustaining
applications.
Intel retains the right to make changes to its test specifications at any time, without notice.
The hardware vendor remains solely responsible for the design, sale and functionality of its product, including any liability arising
from product infringement or product warranty.
Copyright © Intel Corporation 2006. All rights reserved.
Intel, the Intel logo, and EtherExpress are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United
States and other countries.
*Other names or brands may be claimed as the property of others.
Revision 1.4
iii
Table of Contents
Intel® Storage System SSR212MA
Table of Contents
1
Introduction .......................................................................................................................... 9
2
Summary of Issues Fixed in SAN Software 6.6 ................................................................. 9
3
Upgrading to 6.6 ................................................................................................................. 15
Platforms Supported for This Release..................................................................................... 15
3.1
3.1.1
Feature Key Overview ........................................................................................... 15
3.1.2
Prerequisites.......................................................................................................... 15
3.1.3
Schedule Time for The Upgrade............................................................................ 15
3.1.4
Download The Upgrade Components ................................................................... 16
3.2
4
Special Feature Key Upgrade Procedure for This Release ................................... 15
Install the Storage System Software 6.6 Console ................................................. 16
3.2.1
Best Practice.......................................................................................................... 16
3.2.2
Selecting The Type of Upgrade ............................................................................. 17
3.2.3
Verify Management Group Version ....................................................................... 17
Current SAN & SSC Software Limitations........................................................................ 19
4.1
4.1.1
SSC.
4.2
Storage System Console (SSC) ............................................................................ 19
Storage System Console Fails to Install On linux for 6.6.xx.xxxx Release of The
19
Upgrades ............................................................................................................... 20
4.2.1
Upgrade Post-Qualification May Grab Focus Every 20 Seconds .......................... 20
4.2.2
Upgrading Storage Modules and Management Groups May Take Some Time .... 20
4.2.3
Error
On SUSE Linux Enterprise 10, Storage System Console Installation Stops with
20
4.2.4
Storage System Console Will Not Install on Microsoft Windows Longhorn Beta 2
Build 5384 ............................................................................................................................ 21
4.3
4.3.1
Storage System Module ........................................................................................ 21
How to Correctly Identify a Faulty Power Supply................................................... 21
4.3.2
Rebooting the Storage Module While RAID Is Rebuilding Causes Reboot to Take
Up To 20 Minutes ................................................................................................................. 21
4.3.3
4.4
Repair Storage Module Stalls When Attempting To Remove the Storage Module 22
RAID and Disk Management ................................................................................. 23
4.4.1
Why RAID May Go Off If a Foreign Drive Is Inserted Prior To Powering Up the
SSR212MA........................................................................................................................... 23
iv
4.4.2
Swapping One or More Disks across Controllers Causes Data Loss.................... 23
4.4.3
What to Do When A Cache Corruption Alert Is Received...................................... 23
Revision 1.4
Intel® Storage System SSR212MA
4.4.4
Table of Contents
Rebuilding RAID 5 Takes To Long When Minimum Setting is 1 ........................... 24
4.4.5
When Replacing a Disk, if New Disk is seated improperly Disk Status Displays
DMA Off With Yellow Exclamation Icon ............................................................................... 24
4.4.6
Removing Drive from SSR212MA without First Removing Disk from RAID
Requires Rebooting the SSR212MA to Recover from Degraded Mode............................... 24
4.4.7
SSR212MA Becomes Inaccessible after RAID BIOS Deletion.............................. 24
4.4.8
No Warning If Remove and re-Add Disk to RAID 0 ............................................... 26
4.5
Network Management............................................................................................ 27
4.5.1
Storage Traffic Is Using The Wrong Network Interface Card (NIC) ....................... 27
4.5.2
Configuring the SAN on A Private vs. Public Network ........................................... 27
4.5.3
When Jumbo Frames Are Used With Incorrect Network Setup, Management
Group Is Rendered Unusable............................................................................................... 27
4.6
Reporting and SNMP............................................................................................. 28
4.6.1
“NVRAM Card = Corrupt” Alert Generated When the Storage Module is Restarted
After Being Shut Down for Some Time ................................................................................ 28
4.6.2
4.7
NTP and SNMP Setting Are Not Retained After DOM Replacement .................... 28
Management Groups ............................................................................................. 29
4.7.1
Restoring a Management Group Configuration Fails Because the Configuration
Backup Does Not Reflect AN IP Address Change Made To AN SSR212MA ...................... 29
4.7.2
Management Group Restoration Fails When Storage Module Have Different
Software Versions (6.5 and 6.6) And Are Running Managers ............................................. 29
4.7.3
Management Group IP Addresses Are Misconfigured Which Causes Manager
communication Problems When A Storage Module Reboots .............................................. 30
4.7.4
After Deleting A Virtual Manager, Removing A Storage Module From the
Management Gropu Fails The First Time............................................................................. 30
4.8
Clusters.................................................................................................................. 31
4.8.1
If Incorrect Virtual IP Information Is Entered, SSR212MAs Go Offline and Volumes
Become Unavailable ............................................................................................................ 31
4.9
Volumes & Snapshots ........................................................................................... 32
4.9.1
Snapshot Schedules Do Not Adjust For Daylight Savings Time ........................... 32
4.9.2
Creating a Volume with a Duplicate Name Does Not Give an Error...................... 32
4.9.3
Frequent Re-synchronizing Of Snapshots Prevents Other Operations ................. 32
4.9.4
When Running DSM for MPIO, Wait 60 Seconds before Re-logging Onto A
Volume 32
4.9.5
Auto Grow Not Converted to 6.6 Functionality When Upgrading........................... 33
4.9.6
If iSCSI Load Balancing Is Misconfigured, After Upgrading To Version 6.6 Volume
Operations Fail ..................................................................................................................... 33
4.9.7
Revision 1.4
Volume Not Added to Volume List Appears in iSCSI Initiator................................ 33
v
Table of Contents
4.9.8
Intel® Storage System SSR212MA
Cannot Retain Fewer Schedule Remote Snapshots Than Primary Snapshots..... 34
4.9.9
Volume Lists Must Contain Only Authentication Groups with Same Load Balancing
Configurations ...................................................................................................................... 34
4.9.10 Enable or Disable Load Balancing On An Authentication Group Requires Logging
Off And Re-logging On To Volumes ..................................................................................... 34
4.10
Remote IP Copy .................................................................................................... 35
4.10.1 Remote IP Copy from Multiple Management Groups to a Single Remote
Management Group Causes Performance Drop in Remote Management Group................ 35
4.10.2 Remote Copy Schedules Failed To Recreate When Restoring A Remote
Management Group. ............................................................................................................ 35
4.10.3 After Restoration of Primary Management Group, Remote Snapshot Schedule on
Remote Management Group Gets Deleted. ......................................................................... 35
4.11
ISCSI ..................................................................................................................... 37
4.11.1
Adaptec HBA Unable To See Target..................................................................... 37
4.11.2
iSCSI Closes All Shares After Reboot ................................................................... 37
4.11.3 An iSCSI Volume That Becomes Unavailable For Approximately 60 Seconds Or
Longer May Cause Data Loss.............................................................................................. 37
4.11.4 When Mounting Existing iSCSI Volumes On Different Servers, Volumes May Be
Assigned Duplicate Drive Letters Or No Drive Letters ......................................................... 38
4.11.5 Linux-iSCSI Initiator Cannot Reboot When Storage System Software Volume is
Unavailable........................................................................................................................... 38
4.11.6 If Changing Permissions On An iSCSI Volume, Log On To A New Initiator Session
To Complete The Changes .................................................................................................. 38
4.11.7
Microsoft* iSCSI Initiator Does Not Support Dynamic Disks.................................. 38
4.11.8 iSCSI Volume Disappears From The iSCSI Initiator “Active Sessions” Window
When Using Scheduled Snapshots...................................................................................... 38
4.11.9
Unable to Build Oracle Application Cluster on iSCSI Raw Devices....................... 39
4.11.10 SSR212MA Failover Works As Long As A Virtual IP Address Is Used With iSCSI
Initiators from Microsoft*, Intel®, Solaris*, Qlogic*, Adaptec*, Novell*, HP* and IBM* ......... 40
4.11.11 After Power Cycle, Load Balancing Does Not Distribute Requests Properly From A
Microsoft Cluster .................................................................................................................. 41
4.11.12 When Using Storage System Software DSM for MPIO, If User Logs Off A Target
Session While Client Is Accessing A Volume, The Path Disappears From MS Initiator ...... 41
4.11.13 iSCSI Load Balancing Does Not Properly Balance iSCSI Sessions When Running
A Mix of Servers with Storage System Software DSM for MPIO And Servers with iSCSI
Load Balancing Enabled ...................................................................................................... 41
4.11.14 iSCSI Load Balanced Connections Using Virtual IP Are Not Re-Assigned After
Volume Is Migrated or Storage Module Removed from Cluster ........................................... 42
4.11.15 RedHat: Changing Authentication Type Causes Existing iSCSI Devices To Be
Renamed.............................................................................................................................. 42
vi
Revision 1.4
Intel® Storage System SSR212MA
Table of Contents
4.11.16 iSCSI Load Balanced Connections Using Virtual IP (VIP) Are Not Re-assigned
After Volume Is Migrated or Storage Module Removed From Cluster ................................. 43
4.11.17 Failed Initiator Session Records Are Not Always Removed From Database ........ 43
4.11.18 An Extra Microsoft iSCSI Session Is Created In The Console After Rebooting The
Host
43
4.11.19 Microsoft iSCSI Initiator Stops With Error.............................................................. 43
4.11.20 SuSE 9 and SuSE Linux iSCSI: Version 4.0.1-88.26 Initiator Reports Incorrect
Driver State .......................................................................................................................... 44
4.11.21 Periodic iSCSI Event 39 Errors in Windows Host Event Log After Recoverable
SAN Failure .......................................................................................................................... 44
4.11.22 2-way CHAP Does Not Work With Solaris 10........................................................ 44
4.12
Configuration Backup and Restore........................................................................ 45
4.12.1 SSR212MA Post-Install Qualification of Restored SSR212MA Stalls If Restored
SSR212MA Has Different IP Address than That of Original SSR212MA............................. 45
4.12.2 If IP Address on SSR212MA Is Changed Using the Configuration Interface, Some
Processes Continue to Use the Old IP Address................................................................... 45
4.12.3 Single Disk Errors Are Not Recovered In Clusters with SSMs Running Mixed SAN
Software Versions ................................................................................................................ 45
4.13
MCS....................................................................................................................... 47
4.13.1 MCS Cluster Failover While SSR212MA Cluster under Heavy Load Takes MCS
Cluster Off-line ..................................................................................................................... 47
4.14
4.14.1
Revision 1.4
Dell OpenManage Secure Port Server .................................................................. 48
Unable To Install or Load Console with Dell's Secure Port Server Service Started48
vii
Intel® Storage System SSR212MA
Introduction
1 Introduction
The following Release Notes provide information about current limitations in this 6.6.00.2299
release of the Storage System SAN software, and 6.6.00.0099 release of the Storage System
Console (SSC) software for SSR212MA.
2 Summary of Issues Fixed in SAN Software 6.6
Kernel level disk errors are not reported as disk alerts
Menus in the Console sometimes close when scrolling through them
Console open sockets connections can be slow
Unable to right click on the Virtual Manager and remove from management group
The second right click on an object auto-selects an item on the pop-up menu
Incorrect iSCSI Command sequence handling in certain circumstances
Starting and/or stopping a manager could cause an iSCSI disconnect and reconnect
iSCSI VIP failover can take a long time and cause iSCSI disconnect on the server(s)
The BBU display name in Diagnostics panel differs from the Hardware Report and the Alert Panel
Misconfigured Jumbo Frames can cause the management group to lose quorum
Powering off a disk while in a RAID5 configuration can cause an iSCSI disconnect
Re-adding a monitored variable fails with “Trigger entry is null” message
Storage servers and volumes become unavailable when you enter incorrect VIP information for
the cluster
SNMP returns some incorrect value for certain information
Upgrades require that volumes and storage modules finish re-syncing before proceeding
SNMP does not return accurate fan and power supply status
Out-of-sequence Immediate iSCSI commands cause the iSCSI connection to close
Misconfigured Jumbo Frames can cause the management group to be unstable
Console shows the disk status as “Different SSM” during a rebuild when it should be “Active”
When repairing a storage module, all disks are listed at “inactive,” thus not clearly indicating to
you which disk needs repair.
No error message when powering on an active hot spare disk
Changed the ‘Snapshot Status’ monitored variable to ‘Snapshot Schedule Status’ to accurately
reflect what is being monitored by this variable
Able to add a down-version storage module to an up-version management group
SNMP does not return accurate model and power supply information
Changing the NIC bond preference causes the VIP interface to become unavailable if the VIP is
hosted on that storage module
Enhancement: Change the NVRAM cache management to improve reliability
The term “Serial Number” is used to describe what is really the system “MAC Address”
Revision 1.4
9
Summary of Issues Fixed in SAN Software 6.6
Intel® Storage System SSR212MA
Enhancement : Add volume list manipulation support to the CLI
The Administrator can configure iSCSI Load Balancing when there is no VIP on the cluster
Mismatched jumbo frames configuration on storage modules in the same management group can
cause loss of quorum
Enhancement: Implement the Disk IO SNMP MIBs according to UCD-DISKIO-MIB
Too Many manager operations can overload the SAN to the point that volumes become
unresponsive
Support Tools installed on the storage module do not persist through an upgrade
Enhancement: snapshot should no longer inherit volume list membership from their parent volume
Manager leaks small amounts of memory when performing iSNS related tasks
Storage server read-caching algorithm issues non-sequential I/O that can result in some latencies
under workload
Random reads do not scale from queue depth on to queue depth two
Enhancement: Use better NIC bonding names in the NIC setup screens
Monitored Variable dialog email checkbox gets cleared if the email address field is empty and you
click the Finish button
Enhancement: Do not allow a storage module host name to be changed to that of an already
discovered storage module
User can inadvertently start a data restripe using storage module drag-n-drop functionally
Enhancement: Change volume creation defaults to empty for size, hard threshold and soft
threshold
Enhancement: When creating a new authentication group, enable iSCSI by default
Enhancement: Implement local bandwidth limits per Best Practices
Enhancement: Change the New Cluster/Edit Cluster dialog to reduce user-error around iSNS.
Enhancement: Periodically alert you to software upgrade availability
Enhancement: When assigning a volume to a volume list, use Read/Write permission by default
Enhancement: Remove cluster Hot Spare functionality from the Console
Setting the time on the storage module can cause the storage server to become unavailable
Storage server become unavailable if the battery is removed
Enhancement: OS partitions should rebuild RAID at a high priority
Remote copy snapshot retention policies cannot be different on primary and remote volumes
When RAID fails the storage module still hosts iSCSI connections
Enhancement: Add the capability to multi-select items in a list table
Excessive manager communications occur that could slow down the copy progress when remote
copying a sparse snapshot
Enhancement: Display storage module RAID configuration while creating a cluster with the
Cluster Wizard
Erroneous warning dialogs are presented when you use the Reconfigure RAID operation on a
storage module that does not RAID configured
10
Revision 1.4
Intel® Storage System SSR212MA
Summary of Issues Fixed in SAN Software 6.6
Hot-removing a disk from a RAID 0 array causes the manager to become unavailable
Missing or incorrect configuration file can cause some processes on the storage module to restart
Frequent storage server re-starts generate too man log events in the manager log file that can
use the log partition to fill up
The storage server page scrubbing algorithm should be expanded to also check the metadata
contents to avoid meta-data corruption
Enhancement: Remote snapshots should no longer inherit volume list membership from their
parent volume
Volumes created from snapshot rollback do not have the same volume list membership as the
original volume
Storage server journal can overwrite some its entries when the space allotted to the journal is full
Cannot find a storage module if the IP address has a leading space
Powering down a storage module may cause the storage server in another storage module to
become unavailable
Management API Gateway can occasionally become unavailable when unloading idle providers
Storage server become unavailable if the VIP information is invalid
Drop support for management group versions 0 (Storage System Software 4.0) and 1 (Storage
System Software 4.1)
Enhancement: Add multiple volumes to a volume list in a single operation
Enhancement: change the button texts on the Alerts dialog to be more user friendly
When adding more one target volume, VMWare ESX 3.x iSCSI initiator does not correctly
recognize the capacity, the additional targets are not available
Enhancement: Changed the snapshot and remote copy schedule menus to be more intuitive
Managers can become flooded with management group operations, cause the SANI to be
unavailable
Enhancement: When configuring Alerts, default the email checkbox to ON
Graph on the Disk Usage report has clipped text
Management group restore fails with “Unable to set Hot Spare Timeout” message
Enhancement: Warn users if they set the management group local bandwidth above 40 MB/sec
Enhancement: Remove the Hot Spare monitored variable
Enhancement: Add ability to double click the monitored variable list in order to edit to variable
Volume rollback is not retaining original volume hard and soft thresholds
Enhancement: Automatically place text cursor in the appropriate text field in a dialog
CPU Utilization become 100% when hard threshold is reached with auto grow disabled
CANCEL button behaves same as OK button
Enhancement: Make Right-click Menus and Task Menus identical
Enhancement: Enable right-click Copy/Cut/Paste/Delete menus in text fields and areas
Enhancement: Change the name of the Email Setup tab to “Email Server Setup”
Inefficient Manager cleaning operations can reduce restripe performance
Revision 1.4
11
Summary of Issues Fixed in SAN Software 6.6
Intel® Storage System SSR212MA
Enhancement: Change the warning text when a user converts a primary volume to a remote
volume
Enhancement: Changing remote bandwidth defaults to changing the bandwidth symmetrically on
both the Primary and remote groups
Enhancement: Allow user to specify a percentage of the common WAN connection speeds
Changing the storage module host name does not change the storage module name in Alerts
iSCSI session list for a volume is not correct for multiple host NICs.
Reset To Factory Defaults function does not work properly
Console shows incorrect statuses for volumes and storage servers if there is no quorum in the
management group
Inconsistent manager information is displayed in the storage module Details panel
iSCSI MaxCmdSN value I incorrect after a LUN reset is issued from the host
Monitoring software crashes if a Monitored Variable is deleted
Console allows you to setup remote copy to a down-version group
Enhancement: Add multi-select function to volume lists
Enhancement: Display a label on the management group Times tab that indicates the last refresh
time
Enhancement: Change the Test Email button logic to include the Apply function only if needed
Enhancement: Change title from ‘Save Files in directory’ to just ‘Save”
Console attempts to obtain duplicate data numerous times all disks present in the storage module
Sorting the alerts table multiple times yields inconsistent results
Table sorting should be based on complete numbers rather than single digits
Disk temperature alerts may not be generated for all disks that have exceeded the temperature
threshold in a system
Disk SMART Health and Temperature Diagnostics are listed as “Failed” for a missing disk
Enhancement: Auto-fill hard and soft threshold equal to volume size
Enhancement: Add multi-select function to snapshot lists for deletion
Console attempts to obtain disk information repeatedly when RAID is Off
Feature Key error message does not display the storage module host name/IP address for the
storage module in question
The Fan Status Log does not indicate the fan number if a fan is faulty
RAID option “Not Configured” is displayed even when RAID is configured on the storage module
Primary volume gets created with 0MB size while changing volume type from remote to primary
Volume size, hard and soft thresholds are not reset to zero when toggle volume from Primary to
remote
iSCSI session is not re-established if the storage module hosting the session (non-VIP) is
rebooted
Reconfiguring RAID causes the SNMP processes to become unavailable
Remote copy details are not always accurate
12
Revision 1.4
Intel® Storage System SSR212MA
Summary of Issues Fixed in SAN Software 6.6
Present a clear error message to users when they attempt to remove the VIP from a cluster that is
servicing iSCSI load balanced volumes
Cannot set the Replication Priority on a volume when changing Replication Level from Non to 2way to 3-way replication
Storage server is performing an extra memory copy during the copy-on-first-write operation
Selected text in a dialog can sometimes become unselected without a reason
Persistent targets configured in the MS-iSCSI initiator are not connecting reliably after a reboot
Remote copy list is not always sorted correctly
When creating a remote volume, are prompted for size and threshold value when the remote
volume by definition will have zero size and threshold
Persistent targets configured in the MA-iSCSI initiator are reconnected if the volume is moved to a
different cluster
The Test Email function indicates successful delivery to an invalid email address
The Test Email dialog cannot be closed using the “X” button in the upper right corner of the dialog
box
Some remote copy statistics are not always correct
Improve performance of internal data operations like volume migration, restriping, snapshot
deletes, etc
Enhancement: Add SNMP OIDs for cluster Max Used and Max Full statistics
Powering on a disk that was never part of the array results in an unnecessary error message
Erroneous error message may occasionally pop up while creating a management group
Not able to create a volume equal to all remaining cluster space due to rounding error
Enhancement: Make the disabled task menu look disabled
Enhancement: Add a task menu and a right-click option to the Getting Started Page
Remote copy schedules are not suspended if the primary volume becomes a remote volume
Console still uses the IP address from a disabled interface on subsequent storage module
searches
Dialogs are not centered all the time on the underlying Console window
Long labels can get truncated
Setting up mutual CHAP does not work because the initiator secret is lost
Unable to login to a management group from the New Remote Snapshot the New Schedule
Remote Snapshot dialogs
Enhancement: Display the appropriate Stop/Start Virtual Manager menu item based on whether
or not the Virtual Manager is running
Space bar does not activate hyper-linked labels
Improve page scrubbing statistics to know the number of completed scrub cycles and the
completion time of the last scrub cycle
CHAP Authentication fails if the initial byte(s) of the challenge has a leading “0”
Console gathers storage data frequently which interferes with normal Console management
functions
Revision 1.4
13
Summary of Issues Fixed in SAN Software 6.6
Intel® Storage System SSR212MA
User is able to power down a storage module that would result in loss of quorum without any
warning messages
Enhancement: Display partition type of RAID status for all RAID devices in the Hardware
Information Report
Unknown ‘Drive Health alerts are generated if there is an error reading SMART data from the
drives
When removing the storage module from a management group, the storage module status is
showed as “Joining Management Group”
iSCSI IOs stall for ~10 seconds when a snapshot is created
Some system settings (NTP, SNMP, NIC Bonding, etc.) are not reset after you perform a ‘Reset
to Factory Defaults’ via the configuration Interface
The Console command “getting the disk list” take a long time to complete
User is able to remove permanent variables from Monitored Variables list
iSCSI discovery failures with HP iSCSI initiator using iSCSI Digest
Some system settings (NTP, SNMP, NIC Bonding, etc) are not retained after a DOM (Disk-onMemory) replacement
Multiple alerts are sent for BBU Capacity Test Overdue monitored variable
BBU Capacity test kicks off a day late and uses GMT time instead of local time
iSCSI IOs to a volume can stall for ~ 5 seconds when a snapshot is created
Setting the storage module host name can occasionally lead memory when the unit is being
removed from the management group
Repeated alerts with debug messages are received without a time stamp if the storage module is
offline
Enhancement: Storage Module Restore should prevent you from restoring a backup configuration
from a different model storage module
SNMP settings are not restored to the storage module using the backup-restore storage module
function
Volumes list in the remote snapshot dialog is not refreshed after using the login into a remote
management group
14
Revision 1.4
Intel® Storage System SSR212MA
3 Upgrading to 6.6
Platforms Supported for This Release
Upgrades to Release 6.6 are available for SSR212MA. Please use the following procedures to
upgrade storage modules from 6.5.xx.xxxx or 6.3.43.xxxx to Release 6.6.
If you see the following type of message during the upgrade process, please call customer
support.
UPGRDE: 6.6.xx.xxxx.main—Upgrade will now be aborted.
Thu May 19 22:34:56 GMT 2005: UPGRADE: 6.6.xx.xxxx.main aborted
3.1 Special Feature Key Upgrade Procedure for This Release
3.1.1 Feature Key Overview
Storage System Software 6.3 and above
With Storage System Software 6.3 and above, keyed features are enabled per SSM. Feature
keys are required for the following add-on features and applications:
„
„
„
Scalability Pak
Configurable Snapshot Pak
Remote Data Protection Pak
Customers may use these add-on features and applications without a feature key, but are limited
to a 30-day trial period. After the 30-day trial period, if a feature key is not purchased, any
volumes and snapshots associated with add-on features or applications will become unavailable
until a feature key is purchased and applied.
3.1.2 Prerequisites
„
„
„
„
If you are running a software version earlier than 6.3 SP1 (6.3.43.xxxx) on the SSR212MA, you
must first upgrade to 6.3 SP1. You can then upgrade to Release 6.6.
If any iSCSI volumes are in use, stop any activity to those volumes and unmount or log off
before beginning the upgrade. The storage module reboots as part of the upgrade process.
Consequently, your volumes may go off-line, depending on your configuration.
If you are running iSCSI load balancing, you must have Virtual IP Addresses configured.
Ensure that you are using version 6.6.00.0099 of the Console before upgrading the Storage
System Software on the storage modules.
3.1.3 Schedule Time for The Upgrade
Upgrading a storage module from version 6.5.xx.xxxx or 6.3.43.xxxx to 6.6.xx may take 30 to 40
minutes depending upon the specific platform and configuration.
Revision 1.4
15
Upgrading to 6.6
Intel® Storage System SSR212MA
Additionally, after the storage modules are upgraded, rebooted and then found on the network in
the Console, the management group health check may take up to another 10 minutes.
3.1.4 Download The Upgrade Components
1. Download the upgrade components to a temporary location.
2. Download the latest Storage System Console (6.6.00.0099) from the SSR212MA support
site or from your IBL account.
3. Download the appropriate Storage System Software upgrade package(s) from the SSR212
MA support site or from your IBL account.
Platform
SSR212MA
For Storage
System Software
Version
6.3 SP1 or greater
Use Upgrade File
6.6.00.2699.20060929.SSR212MA.upgrade
3.2 Install the Storage System Software 6.6 Console
1.
Install the Storage System Software 6.6.00.0099 Console and discover storage module(s)
on the network.
2.
Use the Console to install the Storage System Software 6.6.xx.xxxx upgrade. If you do not
have a direct path to the 6.6 release, you may be required to first upgrade the software on
the SSM(s) to a version that can then upgrade to 6.6.
3.2.1 Best Practice
ƒ Virtual IP Addresses - If a Virtual IP (VIP) address is assigned to a storage module in a
cluster, the VIP storage module needs to be upgraded last. The VIP storage module is
shown in a field in the clusters detail tab.
ƒ
16
o
First upgrade the non-VIP storage modules that are running managers one at a
time.
o
Then upgrade the non-VIP non-manager storage modules.
o
Lastly, upgrade the VIP storage module.
Remote Copy – If you are upgrading management groups with Remote Copy associations,
you must upgrade the remote management group first. If you upgrade the primary group
first, Remote Copy will stop working.
Revision 1.4
Intel® Storage System SSR212MA
Upgrading to 6.6
3.2.2 Selecting The Type of Upgrade
The Storage System Software Console supports two types of upgrades, as shown in the figure
below.
„ One-at-a-time (recommended) - this is the default and only method if the storage modules exist
in a management group.
„ Simultaneous (advanced) - this allows you to upgrade multiple storage modules at the same
time if they are not in a management group. Use this for new storage modules and/or reconfigured storage modules.
1. Select from the list which storage modules to upgrade.
2. Select the type of upgrade.
3. Click Install.
After The Storage Module Reboots
After each storage module upgrade, during the management group health check, you may
see messages such as “Waiting for MG1 to come up. The issue is –An SSM is down.”
The storage module is not down. It is actually resyncing with the other storage modules in
the management group.
3.2.3 Verify Management Group Version
After the upgrades are complete, the Console attempts to upgrade the management group
version. The rules for this operations are as follows
Revision 1.4
17
Upgrading to 6.6
Intel® Storage System SSR212MA
Upgrade management group version to 6.6 if all storage module serial numbers (eth0 MAC
address) are in the list of known serial numbers included in the upgrade file
„ Do not upgrade the management group version to 6.6 if any of the storage module serial
numbers are unknown.
If you get the following message (or similar message), the management group version upgrade
has not completed. Until the management group is upgraded to 6.6, you will not be able to take
advantage of all the 6.6 features.
„
18
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4 Current SAN & SSC Software Limitations
4.1 Storage System Console (SSC)
4.1.1 Storage System Console Fails to Install On linux for 6.6.xx.xxxx Release of
The SSC.
Issue
When downloading the installer for the Console from the vendor’s FTP site, the FTP program reports that
the download completed succesfully. However, when you run the installer, you receive an error message
indicating that a Java error occurred and the installtion cannot continue.
This occurs because some FTP programs may not download the complete installation package. You can
verify that the download was complete by comparing the MD5 checksum of the file that was downloaded
with the MD5 checksum that is published on the FTP site.
Fix
Upgrade the FTP client you are using or use a different FTP client.
Revision 1.4
19
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.2 Upgrades
4.2.1 Upgrade Post-Qualification May Grab Focus Every 20 Seconds
Issue
During a software upgrade, the Storage System Console may come to the front of other windows open on
the desktop and may grab focus as well.
Workaround
None.
4.2.2 Upgrading Storage Modules and Management Groups May Take Some Time
Issue
Upgrading a storage module from 6.3.xx to 6.6.xx takes from 15-45 minutes depending upon the specific
platform and configuration. Additionally, after the storage modules are upgraded, they have rebooted, and
have all been found on the network in the Console, the management group health check may take up to
another 10 minutes.
Workaround
None.
4.2.3 On SUSE Linux Enterprise 10, Storage System Console Installation Stops
with Error
Issue
You cannot install or run Storage System Console in SLES 10.
Workaround
1. Edit the bin file (CMC_6.6.00.0099_installer_Linux.bin)
2. Comment out the following line
Export LD_ASSUME_KERNEL = 2.2.5
As
#xport LD_ASSUME_KERNEL = 2.2.5
3. Now the installation completes successfully
4. After the installation completes, edit the following file as above.
/opt/LeftHandNetworks/UI/LeftHand_Networks_Centralized_Management_Console
5. Now the Console will start and all the applications will run normally.
20
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.2.4 Storage System Console Will Not Install on Microsoft Windows Longhorn
Beta 2 Build 5384
Solution
Do not attempt to install the Console on Microsoft Windows Longhorn Beta 2 Build.
4.3 Storage System Module
4.3.1 How to Correctly Identify a Faulty Power Supply
The Intel® Storage System SR212MA ships with only one power supply; therefore, only one power supply
is listed in the passive report. Status will be either normal or faulty.
Issue
In a system that has been upgraded to add a redundant module, if a power supply is not working properly,
the storage console passive report it will report power supply status as faulty. The faulty module number will
not be identified.
Workaround
To identify the storage module with a faulty power supply.
1. On the Module Information tab, click Set ID LED On.
2. The ID LED on the left front of the module illuminates a bright blue. Another ID LED is located on
the back of the module on the right side under the empty slot.
3. Go to the back of the storage module and look at the two power supplies.
4. A green LED will be illuminated on the working power supply and an amber LED on the faulty
power supply.
5. Replace the faulty power supply.
Note: To ensure redundancy, the two power cords must be connected to separate and independent
power sources.
4.3.2 Rebooting the Storage Module While RAID Is Rebuilding Causes Reboot to
Take Up To 20 Minutes
Issue
If you reboot the storage module while RAID is rebuilding, the reboot can take up to 20 minutes to
complete.
Cause
The lower the priority setting of the RAID, the longer it will take the reboot to complete.
Revision 1.4
21
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.3.3 Repair Storage Module Stalls When Attempting To Remove the Storage
Module
Workaround
1. Close the Storage System Console and reopen it. The storage module has moved from the cluster
to the management group and the ghost storage module is in the cluster.
2. Remove the storage module from the management group.
Cause
The lower the priority setting of the RAID, the longer it will take the reboot to complete.
22
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.4 RAID and Disk Management
4.4.1 Why RAID May Go Off If a Foreign Drive Is Inserted Prior To Powering Up the
SSR212MA
Issue
If the storage module powers up with a drive that does not belong to the RAID configuration, data
corruption may occur causing RAID to go off and preventing the storage module from coming online.
Replacing the original drive may not result in RAID going to normal. Data may be lost on this storage
module in this case.
Workaround
Drive replacement should ALWAYS be done using the Console. Select the drive to replace, click power-off,
insert a new drive, click power-on, and then click add-to-RAID.
4.4.2 Swapping One or More Disks across Controllers Causes Data Loss
If the storage module powers up with one or more drives foreign to the configuration of a controller, data
corruption occurs.
Issue
The storage module is moved a different physical location. Before the move, the storage module is
powered down and all drives are removed. While replacing the drives back in the drive bays, one or more
drives are accidentally inserted into slots handled by a different controller. When the storage module is
powered up, data corruption occurs.
Workaround
Labels should be added to drive carriers when first installed. If this has not been done, label the drives
before removing them so that you can replace them in the correct bays.
4.4.3 What to Do When A Cache Corruption Alert Is Received
Issue
Cache corruption can occur if the storage module is powered down while there is data in the RAID cache.
If the storage module stays powered-off long enough (more than 72 hours), data in the cache will be
corrupted. When the storage module powers back up, the cache corruption is detected and an alert is
posted indicating the cache is corrupt. The storage module will not be allowed to come online in order to
prevent corruption within the cluster. A “storage module down” alert will also be posted. Please note that
data on the storage module had been lost in the case and must be rebuilt from the cluster assuming
replication was configured.
Workaround
To resolve this issue, please contact support
Revision 1.4
23
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.4.4 Rebuilding RAID 5 Takes To Long When Minimum Setting is 1
Issue
The default setting for the minimum RAID rebuild rate is 1. This setting may cause RAID 5 rebuild to take
too long.
Workaround
Increase the minimum rebuild rate to a value of 10 or greater. The following guidelines describe the effects
of the RAID rebuild rates.
• Setting the rate high is good for rebuilding RAID quickly and protecting data; however, it will slow
down user access.
• Setting the rate low maintains user access to data during the rebuild.
4.4.5 When Replacing a Disk, if New Disk is seated improperly Disk Status
Displays DMA Off With Yellow Exclamation Icon
Issue
A disk is replaced in an SSR212MA. After the RAID rebuild is complete, the disk status displays DMA Off.
This status occurs due to an improperly seated disk.
Workaround
Repeat the procedures for replacing the disk, paying careful attention to reseat the disk properly in the drive
bay. After the RAID rebuild is finished, the disk status should be correct.
4.4.6 Removing Drive from SSR212MA without First Removing Disk from RAID
Requires Rebooting the SSR212MA to Recover from Degraded Mode
Issue
If a drive is removed without first removing it from RAID in the Console, RAID becomes degraded and the
SSR212MA becomes inaccessible.
Workaround
1. Re-insert the drive.
2. Reboot the module.
3. Add the disk to RAID. RAID will start rebuilding after the drive is powered on.
4.4.7 SSR212MA Becomes Inaccessible after RAID BIOS Deletion
Issue
After deleting RAID configuration from the RAID BIOS in the MA system, the system can no longer be
accessed, both through SSC connection or through the serial link. The system will boot up normally and
responde to pings from the network, however, cannot be access through the SSC software. When trying to
connect through the serial link you will get error: “Error opening a connection”.
Workaround
24
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
There is currently no workaround for this issue.
Revision 1.4
25
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.4.8 No Warning If Remove and re-Add Disk to RAID 0
Issue
SSR212MA is configured with RAID 0. While the SSR212MA is running, user manually removes any disk
from the SSR212MA. On the Disk Setup window the disk status is “Off or missing.” On the RAID Setup
window, RAID status is Normal.
This Issue occurs when the disk is removed while there is no activity to the volume. As soon as any activity
to that volume occurs, such as a client attempting to read or write data, then the volume becomes
unavailable.
26
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.5 Network Management
4.5.1 Storage Traffic Is Using The Wrong Network Interface Card (NIC)
Issue
You may see Storage System Software traffic on NICs other than the designated one. This is unavoidable
when two or more NICs are assigned IP addresses in the same subnet. It can occur in any configuration
where hosts are configured with multiple NICs.
Workaround
Assign “public” adapters, intended for servicing users, to a subnet distinct from storage adapters.
4.5.2 Configuring the SAN on A Private vs. Public Network
Issue
The recommended best practice is to isolate the SAN, including Console traffic, on a separate network. If
the SAN must run on a public network, use a VPN to secure data and Console traffic.
Workaround
None.
4.5.3 When Jumbo Frames Are Used With Incorrect Network Setup, Management
Group Is Rendered Unusable
Issue
If a management group is created with a storage module that has an improper network configuration, such
as an active-backup bond with one network interface connected to a 10/100 switch and another connected
to a GigE switch, the management group losses connectivity.
Also, if such a bond is deleted, the storage module permanently loses connectivity.
Workaround
A proper network configuration is required when creating a bond; that is, both network interfaces must be
connected to appropriate switches. For more information, see the Intel® Storage System Software User
Manual, “Configuring NIC Bonding.”
Revision 1.4
27
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.6 Reporting and SNMP
4.6.1 “NVRAM Card = Corrupt” Alert Generated When the Storage Module is
Restarted After Being Shut Down for Some Time
Workaround
To resolve this issue, please contact support.
4.6.2 NTP and SNMP Setting Are Not Retained After DOM Replacement
Workaround
Reconfigure the NTP and SNMP settings in the Storage System Console.
28
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.7 Management Groups
4.7.1 Restoring a Management Group Configuration Fails Because the
Configuration Backup Does Not Reflect AN IP Address Change Made To AN
SSR212MA
Issue
This Issue can occur when the IP address change is applied to an SSR212MA, but does not get stored in
the system before the management group configuration is backed up.
Workaround
If you do restore a management group configuration with an incorrect IP address, take the following steps:
1. Log in to the SSR212MA with the incorrect IP address and change it back to the IP address
stored in the configuration backup file.
2. Complete the management group restoration.
3. When the management group is successfully restored, change the IP address of the SSR212MA
to the desired address.
Best Practice to Prevent This Problem
1. In the Edit Configuration window, change the IP address of the SSR212MA.
2. In the Console Network View, select the management group.
3. Right-click and select Backup Configuration of Management Group.
4. In the Back up Configuration window that opens, scroll down to the section labeled SSR212MAs.
5. In the SSR212MAs section, verify that the Communication IP is the new IP address. If it is not the
new IP address, then click OK to cancel out of the Backup Configuration window.
Wait for a few minutes and then repeat steps 3 through 5. When the correct new IP address appears, select
Back up Configuration. The management group configuration is backed up with the correct IP address.
4.7.2 Management Group Restoration Fails When Storage Module Have Different
Software Versions (6.5 and 6.6) And Are Running Managers
Solution
1. Upgrade all the storage modules to the same version of the software.
2. Backup the management group again, after all storage modules have been upgraded. Then
future management group restorations will work.
Or
1. On storage modules running version 6.5, you can start and stop managers (as long as quorum
can be maintained) so that only the 6.5 storage modules in the management group are
managers.
2. Backup the management group configuration again and restore.
The key to implementing either solution is to take a backup of the management group configuration again
before attempting to restore.
Revision 1.4
29
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.7.3 Management Group IP Addresses Are Misconfigured Which Causes Manager
communication Problems When A Storage Module Reboots
Issue
Creating a management group using the wizard results in an incorrectly configured unicast list. When a
storage module in the management group reboots, the management group may seem to have lost quorum.
Workaround
For each storage module in the management group:
1. Open the Edit Configuration window.
2. Select the TCP/IP configuration category and select the Communication tab.
3. Click Update (at the bottom right corner)
4. Close the Edit Configuration window.
4.7.4 After Deleting A Virtual Manager, Removing A Storage Module From the
Management Group Fails The First Time
Issue
You attempt to remove a storage module from a management group. The storage module is running a
manager. The management group has a Virtual Manager added and started. First you stop the Virtual
Manager. Then you delete the Virtual Manager from the management group. Next you stop the manager
on the storage module. Now, when you remove the storage module from the management group, the
process seems to complete, but the storage module does not actually get removed.
Workaround
Perform the remove operation again on the storage module. The straoge module should be removed from
the management group.
30
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.8 Clusters
4.8.1 If Incorrect Virtual IP Information Is Entered, SSR212MAs Go Offline and
Volumes Become Unavailable
Issue
When configuring VIP for a cluster, entering incorrect information for any of the components (IP Address,
Subnet Mask and Default Gateway) causes the SSR212MAs in that cluster to go down and any volumes
associated with the cluster to become unavailable.
Workaround
The iSCSI VIP must be in the same subnet as all the SSR212MAs in the cluster.
1. Enter the correct information for the Virtual IP configuration.
2. Reboot the SSR212MAs in the cluster.
.
Revision 1.4
31
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.9 Volumes & Snapshots
4.9.1 Snapshot Schedules Do Not Adjust For Daylight Savings Time
Issue
When snapshot schedules are created under Standard Time, the schedules continue to execute at the
originally scheduled Standard Time, even though the storage modules are operating under Daylight
Savings Time.
For example, if a schedule is configured under Standard Time to run at 2:00 PM, then the schedule initially
runs (under Standard Time) at 2:00 PM. Then, when the local time changes to Daylight Savings Time, the
schedule starts running at 3:00 PM instead of 2:00 PM. This is happening because the schedule is
operating as if Daylight Savings Time doesn't exist; so the schedule continues to execute at 2:00 PM
Standard Time. The Storage System Software does not include automatic adjustments for Daylight Savings
Time.
Workaround
If you want snapshot schedules to operate at the same relative time all year, you must manually edit the
schedules when the time changes in the spring and autumn.
4.9.2 Creating a Volume with a Duplicate Name Does Not Give an Error
Issue
You create a volume in a cluster named Volume_0. You then create a second volume in that cluster named
Volume_0. No error message is generated, the second volume does not appear, but some settings on the
first volume are changed to those entered for the second volume.
Workaround
Do not use the same name for a volume more than once in a cluster.
4.9.3 Frequent Re-synchronizing Of Snapshots Prevents Other Operations
Issue
Re-synchronizing snapshots occupies system resources and may prevent other management group
operations.
Best Practice
Create snapshot schedules with a frequency greater than one hour.
4.9.4 When Running DSM for MPIO, Wait 60 Seconds before Re-logging Onto A
Volume
Issue
32
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
If you are running Storage System Software DSM for MPIO and you log off a volume and immediately log
back on the volume, you may find that the volume does not appear on the Disk Management tab of the
Storage configuration category.
Workaround
After logging off a volume while running Storage System Software DSM for MPIO, wait for about 60
seconds before logging back on to the volume.
4.9.5 Auto Grow Not Converted to 6.6 Functionality When Upgrading
Issue
Auto grow has changed from version 6.5 to version 6.6. When you upgrade, volume configured with auto
grow is set to Enabled in version 6.6. You cannot tell if it is using the manual algorithm or automatic auto
grow.
Workaround
•
Use the scripting command to determine the auto grow setting.
Volume_autogrow_get <volume name> [ <failure timeout seconds>]
See the scripting chapter in the SAN user manual for more information about using scripting.
•
You can reset auto grow on the volume if necessary. Simply use the Console to disable auto grow
and reenable auto grow.
4.9.6 If iSCSI Load Balancing Is Misconfigured, After Upgrading To Version 6.6
Volume Operations Fail
Issue
After upgrading from version 6.5 to version 6.6, various volume operations fail. Snapshot schedules appear
to stall, iSCSI log ins do not work, etc. The following error message displays:
“A Virtual IP is required. A VIP is required when linking or moving components to a load-balancing
authentication group. Edit Cluster to enable the VIP.”
Solution
Before upgrading, ensure that any clusters containing volumes in load-balanced authentication groups have
a VIP configured. Clearing the load-balance flag on the authentication group(s) or configuration a VIP on
the cluster(s) before the upgrade prevents the problem.
Workaround
However, you have already upgraded, there are a couple of solutions to try:
•
Configure a VIP on the cluster that is causing the problem.
•
If there is only one load-balanced authentication groups, clearing the load balance flag will fix the
problem.
4.9.7 Volume Not Added to Volume List Appears in iSCSI Initiator
Issue
Revision 1.4
33
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
You create a cluster and configure the cluster to use iSNS. You then create a volume but do not add the
volume to a volume list. The volume appears as a target in the iSCSI initiator. However, if you attempt to
log on to this target, you receive an Authorization Failure message. This is a function of iSCSI discovery.
Solution
If you need to log on to the volume, add it to a volume list and create an authentication group, as described
in the user documentation.
4.9.8 Cannot Retain Fewer Schedule Remote Snapshots Than Primary Snapshots
Issue
Configure a remote snapshot schedule with a retention policy of 7 primary snapshots and 4 remote
snapshots. The SAN retains 7 primary and 7 remote snapshots
Cause
You cannot retain fewer remote snapshots than primary snapshots. However, you can retain more remote
snapshots than primary snapshots.
4.9.9 Volume Lists Must Contain Only Authentication Groups with Same Load
Balancing Configurations
Issue
If a volume list contains one authentication group with load balancing and one authentication group without
load balancing, it may not be possible for both of two cluster iSCSI clients to connect to the volume at the
same time.
Solution
Only add authentication groups with the same load balancing configuration to a volume list.
4.9.10 Enable or Disable Load Balancing On An Authentication Group Requires
Logging Off And Re-logging On To Volumes
Issue
The user changes the Enabled Load Balancing confugration of an authentication group. After-wards, some
iSCSI clients may not ber able to reconnect to volumes because of the changes.
Workaround
Log off all iSCSI connections and log back on to reset the connections properly.
34
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.10 Remote IP Copy
4.10.1 Remote IP Copy from Multiple Management Groups to a Single Remote
Management Group Causes Performance Drop in Remote Management
Group
Issue
A remote management group experiences a performance drop if too much bandwidth is used for transfer of
Remote IP Copy data.
Workaround
To designate enough bandwidth for I/O to the management group, reduce the bandwidth used for Remote
IP Copy.
1. Log in to the remote management group.
2. On the Edit Remote Bandwidth dialog window, reduce the remote bandwidth setting.
4.10.2 Remote Copy Schedules Failed To Recreate When Restoring A Remote
Management Group.
Issue
Two management groups, A and B, had a remote copy schedule copying from A to B. Both management
group configurations had been backed up.
1. Management group A goes down and is restored from the configuration backup.
2. Later, management group B goes down and is restored from backup but now the remote
schedule is lost.
Workaround
After restoring management group a, back up both management group configurations again. Otherwise you
must manually re-enter the remote copy schedule.
4.10.3 After Restoration of Primary Management Group, Remote Snapshot
Schedule on Remote Management Group Gets Deleted.
Issue
After performing a backup and restore of a primary management group with remote snapshots and remote
snapshot schedules, the remote snapshot schedules on the remote management group are not properly
restored. The schedule on the primary side still exists but is faulty, as if has not corresponding schedule on
the remote side. Schedule remote copies will fail to start.
This failure to properly restore the schedule appears to happen in two situations.
•
When there is a delayed start on the remote snapshot schedule. For example, when the remote
snapshot schedule is setup, the initial start time is set at some point in the future, such as the next
day
•
When there are multiple remote snapshot schedules to restore.
Workaround
Revision 1.4
35
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
Delete the old remote snapshot schedules on the primary management group and recreate them.
Scheduled remote snapshots will then resume.
36
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.11 ISCSI
4.11.1 Adaptec HBA Unable To See Target
Workaround
Do not use MS iSCSI initiator with the Adaptec HBA
4.11.2 iSCSI Closes All Shares After Reboot
Issue
If your iSCSI volumes are used by automatically-started Windows services (e.g., File Sharing), you must
use the Microsoft* Initiator's “Bind Volumes” operation to make sure that those volumes are available before
the services that require them are started.
Workaround
See the Microsoft* support article 870964 on the Microsoft support web site.
Also, see the section entitled “Running automatic start services on iSCSI disks” in the Microsoft* iSCSI
Initiator Users Guide for more details.
4.11.3 An iSCSI Volume That Becomes Unavailable For Approximately 60 Seconds
Or Longer May Cause Data Loss
The Windows Registry has a default maximum hold time setting of 60 seconds before a Microsoft
Windows* system terminates a connection to an iSCSI device that is unavailable.
Therefore, an iSCSI volume that becomes unavailable for longer than 60 seconds may cause delayed write
failures and potential data loss.
Solution
Change the Windows Registry setting for the default Maximum Request Hold Time to a very large (infinite)
value.
Important: Back up your registry before making any changes.
1. Run regedit.exe.
2. Navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\4D36E97BE325-11CE-BFC1-08002BE10318\####\Parameters (where #### is the index of the Microsoft*
iSCSI initiator in the set of SCSI and RAID Controllers).
3. Double-click MaxRequestHoldTime in the right-hand pane. The Edit DMWord Value window
opens.
4. Change the Base to decimal.
5. Enter a value of 600.
6. Click OK.
7. Save your changes and exit the Registry.
8. Reboot the system.
Revision 1.4
37
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.11.4 When Mounting Existing iSCSI Volumes On Different Servers, Volumes May
Be Assigned Duplicate Drive Letters Or No Drive Letters
Issue
An iSCSI volume that was mounted on a server and assigned a drive letter is logged off from Server 1. It is
then mounted on Server 2. Sometimes it picks up a drive letter that is already in use on Server 2.
Sometimes it is not assigned a drive letter. The volume then becomes inaccessible.
Workaround
Open the Windows Disk Management console and assign a new drive letter to the volume. The volume
should then appear in the directory structure.
4.11.5 Linux-iSCSI Initiator Cannot Reboot When Storage System Software Volume
is Unavailable
The iSCSI Device Manager hangs when network problems prevent it from communicating with an
SSR212MA. Because the default timeout for the Linux-iSCSI initiator is infinite, the initiator cannot reboot
when it is unable to access the iSCSI volume on the SSR212MA.Workaround
Restore full network connectivity between iSCSI initiators and SSR212MAs. If this is not possible,
disconnect the SSR212MA that the initiator can't communicate with from the network. Disconnecting will
cause the managers to tell the client that it should stop attempting to contact that SSR212MA.
4.11.6 If Changing Permissions On An iSCSI Volume, Log On To A New Initiator
Session To Complete The Changes
Issue
An iSCSI volume is mounted as a read/write volume and is in use.
You change the access permissions to read-only for the authentication group in the Console.
The permissions have not changed for the clients that are accessing the volume. They are still able to write
to the volume.
Solution
To complete the process of changing permissions, you must log off the current initiator session for that
volume and log on to a new session.
4.11.7 Microsoft* iSCSI Initiator Does Not Support Dynamic Disks
Issue
The Microsoft iSCSI initiator software does not support dynamic disks.
Workaround
Do not create dynamic disks to be used with the Microsoft iSCSI initiator.
4.11.8 iSCSI Volume Disappears From The iSCSI Initiator “Active Sessions”
Window When Using Scheduled Snapshots
Issue
38
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
If you are using scheduled snapshots with an iSCSI volume, and the snapshot hard threshold is set to less
than the volume hard threshold, the iSCSI volume disappears from the initiator Active Sessions window
when the snapshot hard threshold is exceeded.
To recover from this situation:
1. In the Console, edit the snapshot schedule to increase the hard threshold.
2. Re-log in to the volume in the iSCSI initiator.
Workarounds
1. In the snapshot schedule, set the snapshot hard threshold to the same value as the volume
thresholds, or
2. Use the auto_grow scripting feature to configure automatic threshold increases for the volume
hard thresholds.
4.11.9 Unable to Build Oracle Application Cluster on iSCSI Raw Devices
Issue
If more than one iSCSI initiator attempts to access the same volume, the file system on the volume may
become corrupted.
Workaround
Do not use volumes in a clustered iSCSI node configuration. Configure one iSCSI initiator per volume.
Revision 1.4
39
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.11.10
SSR212MA Failover Works As Long As A Virtual IP Address Is Used
With iSCSI Initiators from Microsoft*, Intel®, Solaris*, Qlogic*, Adaptec*,
Novell*, HP* and IBM*
Issue
To take advantage of the Storage System Software failover functionality in the iSCSI initiators from the
listed companies, use a Virtual IP address when configuring clusters in the Console.
Workaround
Adaptec* Initiator
To ensure iSCSI volume availability in case of failover, it is recommended that the Session Recovery
Timeout be set to 600 seconds. This is done using the Adaptec* iConfig utility.Qlogic*Initiator
To ensure iSCSI volume availability in case of failover, the following initiator configuration parameters must
be set via the “Config Parameters” button on the Target Settings tab:
1. Default Timeout: 600 seconds
2. Connection Keep Alive Timeout: 600 seconds
Microsoft* Initiator
To ensure iSCSI volume availability in case of failover, the following registry key must be set:
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E97B-E32511CE-BFC1-08002BE10318}\0000\Parameters]
MaxRequestHoldTime: 600 seconds
Note: The “0000” before “Parameters” in the registry path might vary. It could be 0001, 0002, etc. Search
for MaxRequestHoldTime to find the key.
40
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.11.11
After Power Cycle, Load Balancing Does Not Distribute Requests
Properly From A Microsoft Cluster
Issue
A Storage module is powered off and then powered on, and another storage module in the Storage System
Software cluster handles all the connections to the volumes connected to that cluster. When the storage
module is powered on again, load balancing does not redirect I/O to that storage module.
Workaround
1. Take one of the MS Cluster group offline.
2. Disconnect the iSCSI connection on both storage modules.
3. Reconnect the targets on both storage modules.
4. Bring the MS Cluster group back online.
5. Repeat steps 1-4 for all MS Cluster groups that host Storage System Software iSCSI disks.
4.11.12
When Using Storage System Software DSM for MPIO, If User Logs Off
A Target Session While Client Is Accessing A Volume, The Path Disappears
From MS Initiator
Issue
In the MS iSCSI initiator, when you open the Details window from the Targets tab and click on the
Connections button on the Sessions tab, the number of iSCSI sessions displayed for a given volume is
fewer than expected.
Workaround
1. Quiesce Client activity to the volume.
2. Completely log off the volume and reconnect.
3.
4.11.13
iSCSI Load Balancing Does Not Properly Balance iSCSI Sessions
When Running A Mix of Servers with Storage System Software DSM for
MPIO And Servers with iSCSI Load Balancing Enabled
Issue
A mixture of servers with Storage System Software with DSM for MPIO (Server Group-1) and servers with
iSCSI Load Balancing enabled (Server Group-2) are accessing volumes in a storage cluster. iSCSI
sessions from Server Group-2 are not properly load-balanced.
Workaround
If practical, the problem can be avoided by partitioning the Management Group into storage clusters such
that a given cluster is accessed by only DSM hosts or only –DSM hosts.
Revision 1.4
41
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.11.14
iSCSI Load Balanced Connections Using Virtual IP Are Not ReAssigned After Volume Is Migrated or Storage Module Removed from Cluster
Issue
After Migrating a volume or removing a storage module from a cluster, load-balanced iSCSI sessions are
not re-assigned. While iSCSI connectivity is maintained throughout these operations, performance may not
be optimal.
Workaround
1) Quiesce client activity to the volume.
2) Completely log off the volume and reconnect.
4.11.15
RedHat: Changing Authentication Type Causes Existing iSCSI Devices
To Be Renamed
Issue
You configured an authentication group for iSCSI access. You then changed the access configuration,
either to require CHAP or to remove or change CHAP requirements. After the change, the existing iSCSI
devices are renamed and cannot be remounted.
Workaround
To change the authentication type of any volume (LVM or otherwise)
1. Unmount volumes and stop iSCSI services
# /etc/init.d/iscsi stop
2. Make appropriate changes to the authentication group (i.e. change from iqn to CHAP)
3. Make appropriate changes to the initiator (i.e. settings in /etc/iscsi.conf)
4. Start iSCSI services and remount volumes.
For LVM volume groups, the following steps are recommended since the system allows iSCSI services to
be stopped even though iscsi_sfnet driver is still in use by the volume group.
To change authentication type of volumes being used in a volume group
1. Unmount volume/volume group
# umount /iSCSI
2. Deactivate the volume group
# vgchange –a n vgiSCSI
3. Stop iSCSI services
# /etc/init.d/iscsi stop
Then change to use CHAP or whatever authentication you want to test next. Then restart things in the
reverse order:
# /etc/init.d/iscsi start
# vgchange –a y vgiSCSI
42
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
# mount/dev/vgiSCSI/lvol0/iSCSI
4.11.16
iSCSI Load Balanced Connections Using Virtual IP (VIP) Are Not Reassigned After Volume Is Migrated or Storage Module Removed From Cluster
Issue
After migrating a volume or removing a storage module from a cluster, load-balanced iSCSI sessions are
not re-assigned. While iSCSI connectivity is maintained throughout these operations, performance may not
be optimal.
Workaround
1. Quiesce client activity to the volume.
2. Completely log off the volume and reconnect.
4.11.17
Failed Initiator Session Records Are Not Always Removed From
Database
Issue
When a host disconnects ungracefully from a Storage System Software volume (e.g. the network fails, the
host hardware is rebooted), the Console shows the iSCSI session for that host as Failed. If the host does
not reestablish the connection within a day, the session is supposed to be considered dead and removed
from the Console’s iSCSI session displays. However, the failed sessions are usually not removed, resulting
in lists of failed sessions displaying in the Console.
Workaround
None, However, as long as your volume is connected and accessible, and shows a connected session in
the Console, you can ignore the failed session.
4.11.18
An Extra Microsoft iSCSI Session Is Created In The Console After
Rebooting The Host
Issue
An extra iSCSI session is created in the Console after rebooting the host for the volume which is mounted
with “Automatically restore this connection when the system boots” selected.
Explanation
This is a Microsoft issue in which different session IDs (iSCSI ISIDs) are used for the same host-volume
pair, depending on how the session was established. After an ungraceful host shutdown, you might see
duplicate iSCSI sessions in the Console, one with a Status of Failed and one a Status of Connected.
Workaround
Log off the automatically logged on persistent session and manually log back on to get rid of the spurious
session.
4.11.19
Microsoft iSCSI Initiator Stops With Error
Explanation
In rare cases, the Microsoft iSCSI Initiator version 2.02 and 2.03 may stop after a storage module reboots.
Revision 1.4
43
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
Workaround
Manually restart the Microsoft iSCSI Initiator Service.
4.11.20
SuSE 9 and SuSE Linux iSCSI: Version 4.0.1-88.26 Initiator Reports
Incorrect Driver State
Solution
Use the iSCSI initiator provided with the SLES 9 distribution.
4.11.21
Periodic iSCSI Event 39 Errors in Windows Host Event Log After
Recoverable SAN Failure
Issue
Windows Event logs show long sequences of iScsiPrt event id 39 (task management commands sent)
every 30 seconds accompanied by iScsiPrt event id 27 (no match for tag) and event id 9 (target did not
respond in time) while IO is flowing to a volume. This could indicate that the MS iSCSI session is in the
state where it issues spurious LUN resets.
Workaround
1) Quiesce the application that is using the volume
2) Log off the volume and log back on
3) Restart the application
Or
Reboot the application server
4.11.22
2-way CHAP Does Not Work With Solaris 10
Issue
Volume associated with an authentication group configured for 2-way CHAP cannot be mounted on Solaris
10.
Workaround
Use 1-way CHAP or no CHAP with Solaris !0.
44
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.12 Configuration Backup and Restore
4.12.1 SSR212MA Post-Install Qualification of Restored SSR212MA Stalls If
Restored SSR212MA Has Different IP Address than That of Original
SSR212MA
Issue
Back up an SSR212MA configuration file (SSR212MA-1). SSR212MA-1 becomes unavailable and you
restore the backed up configuration of SSR212MA-1 to a second SSR212MA on the network (SSR212MA2). SSR212MA-2 has a different IP address than the unavailable SSR212MA-1. As part of the post-install
qualification, the Console searches for the newly configured SSR212MA-2 on the network. However, it is
searching for the original IP address of SSR212MA-2 instead of the IP address that was saved in the
SSR212MA-1 configuration back-up file. That search never completes because the IP address on
SSR212MA-2 has changed and is now the IP address of SSR212MA-1.
Note: Restoring multiple SSR212MAs from a single backup file causes an IP address conflict.
Solution
Before restoring a backed-up SSR212MA configuration file, make certain that the new SSR212MA is
configured with the IP address of the original SSR212MA.
Workaround
If the backed up configuration has been restored and the post-install qualification process can't complete
because it cannot find the SSR212MA on the network, do the following:
1. On the Post install qualification window, click Cancel All Installs.
2. Search for the SSR212MA on the network using the correct IP address, or Find by Subnet and
Mask.
4.12.2 If IP Address on SSR212MA Is Changed Using the Configuration Interface,
Some Processes Continue to Use the Old IP Address
Issue
An SSR212MA in a management group has an IP address assigned. That IP address is changed using the
Configuration Interface instead of using the Console. The new IP address is not universally updated in the
Storage System Software and some functions continue to use the old IP address.
Workaround
To finish updating the IP address using the Console:
1. Log in to the SSR212MA with the new IP address.
2. In the SSR212MA Configuration Interface, navigate to the TCP/IP Network category.
On the Communication tab, click Update to synchronize the IP addresses of all managers.
4.12.3 Single Disk Errors Are Not Recovered In Clusters with SSMs Running Mixed
SAN Software Versions
Issue
Revision 1.4
45
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
Release 6.3 contains functionality to recover from any single disk unrecoverable data error. This recovery
functionality only works on SSR212MA’s in clusters where all SSR212MA’s are upgraded to version 6.3. If
a cluster has one or more SSR212MA’s running an earlier version of the software, than the recovery
functionality will not work.
Workaround
Upgrade all SSM’s to release 6.3 SAN software.
Fix
None.
46
Revision 1.4
Intel® Storage System SSR212MA
Current SAN & SSC Software Limitations
4.13 MCS
4.13.1 MCS Cluster Failover While SSR212MA Cluster under Heavy Load Takes
MCS Cluster Off-line
Issue
If an MCS cluster failover occurs while the SSR212MA cluster is under very heavy load, the MCS cluster
does not come back online until the load on the SSR212MA cluster decreases.
Workaround
Increase the “pending timeout” of each of the disk resources on the MCS cluster to the same as the
“maxrequestholdtime” of 600.
Do the following on each “physical disk” resource that is actually an iSCSI disk on the SSR212MA.
1. Right-click on the disk in the MCS cluster administrator.
2. Select Properties > Advanced tab.
3. Change the “pending timeout: seconds” from 180 to whatever you used as a
“maxrequestholdtime” for iSCSI in the registry.
Revision 1.4
47
Current SAN & SSC Software Limitations
Intel® Storage System SSR212MA
4.14 Dell OpenManage Secure Port Server
4.14.1 Unable To Install or Load Console with Dell's Secure Port Server Service
Started
Issue
Using Microsoft Windows* on a Dell* Server with the Dell* OpenManage Secure Port Server service, the
user cannot properly install the Console or start the Console.
Workaround
Stop the Dell* OpenManage Secure Port Server service when installing or running the Console.
48
Revision 1.4
Download PDF