VERITAS Volume Manager 2.7 for Windows 2000 HotFix04 Japanese Version

Article:TECH18592  |  Created: 2002-01-25  |  Updated: 2004-01-28  |  Article URL http://www.symantec.com/docs/TECH18592
Article Type
Technical Solution

Environment

Issue



VERITAS Volume Manager 2.7 for Windows 2000 HotFix04 Japanese Version

Solution



*NOTE* THE ATTACHED FILE IS SPECIFICALLY FOR THE JAPANESE VERSION OF VOLUME MANAGER!
The English version of this hotfix may be obtained through the "Related Document" link provided at the end of this document.

February 21, 2002

This Readme file has the following topics:

I. INSTALLING THE HOTFIX
II. UNINSTALLING THE HOTFIX
III. ISSUES RESOLVED IN THIS HOTFIX

I. INSTALLING THE HOTFIX

If you have a cluster, install the hotfix on each node of the cluster. See the note below for details. Use the following steps to install this hotfix:

1. On the VERITAS Support Web site, access the product listing for Volume Manager for Windows 2000.
2. Under "Recently Published Patches and Updates," click to select the appropriate link for this hot fix.
The file name includes VM2K27HF, the number of the hotfix, and ENU for English or JPN for Japanese.
3. Scroll to the bottom of the Readme file and click Download Now.
4. When prompted, select "Save This Program To Disk" and click OK.
5. At the next screen, select a location to save the file and click Save.
6. After the download completes, click Open to open the VERITAS Self-Extractor.
7. Select a location to save the extracted files and then extract the contents by clicking the extract button. The hotfix package contains three files -- a package ID, this Readme file, and the hotfix executable file, setup.exe.
8. Click OK at the message "All files have been extracted." This Readme file will open automatically.
9. Copy the hotfix executable file to any directory on the computer where Volume Manager is installed.
10. Double-click the hotfix executable file to start the installation. The install utility Welcome screen appears.
11. Click Next to continue.
12. Click Install to proceed with the installation. The hotfix installation runs automatically and will install the files modified by the hotfix to the appropriate directories.
13. In the final screen, click Finish. A message appears, prompting you to restart your system.
14. Click Yes to restart the computer now or No if you plan to restart later. You must reboot to have the changes take effect.

Note: The first time Volume Manager 2.7 is launched after installing this hotfix, you may see several pop-up windows with progress bars while files are loading.

Note: A cluster does not have to be brought down to install the hotfix. A rolling upgrade can be done so that the cluster resources remain available during the installation (see steps a through d below).

a. Before installing the hotfix to a cluster node, move all cluster resources to another cluster node.
b. Install the hotfix as directed above.
c. Move the cluster resources back to the original node.
d. Repeat steps a through c for each remaining node of the cluster.

II. UNINSTALLING THE HOTFIX

Normally you would not need to uninstall the hotfix; but if it becomes necessary to do so, here are the steps:

1. Open the Control Panel and select Add/Remove Programs.
2. Click to select the VERITAS Volume Manager 2.7 for Windows 2000-Hot Fix 04 and then click Remove.
3. Click Yes to respond to the confirmation box that appears. After the process for uninstalling is complete, a message appears, prompting you to restart your system.
4. Click Yes to restart the computer now or No if you plan to restart later. You must reboot to have the changes take effect. The process for uninstalling is now complete.

Note: If you want to repair the hotfix, use steps 1 and 2 above but select Change instead of Remove. A dialog box appears with two options -- Repair and Remove. Select Repair and then click Next. The hotfix repair option runs automatically. It will fix missing or corrupt files, shortcuts, and registry entries.

III. ISSUES RESOLVED IN THIS HOTFIX

This hotfix was created to address several current issues. Because it is a cumulative hotfix, it also contains fixes for problems in all previous hotfixes. This Issues section is divided into two topics:

- Current Issues for This Hotfix
- Past Issues from Previous Hotfixes

A. Current Issues for This Hotfix

1. Resolved several intermittent problems that occurred when MSCS resources were in the process of failover or when resources went offline, online, or were terminated.

a. When a resource is offlined or terminated, error 4 (can't stop SCSI reservation thread) occurs. The resource cannot be failed over or brought back online (Incident 76870).
b. During rapid failover of MSCS resources on a multiprocessor server, bugcheck 0xOa can occur (Incident 76871).
c. When multiple resources are offlined and onlined simultaneously, the Volume Manager application can hang. (Incident 76877).
d. When resources are being failed over continuously, the Volume Manager application can AV (Access Violation). (Incident 76878).

2. Improved the speed of the failover process (Incident 76879).

3. When an MSCS cluster node is started or cluster resources are moved from one node to another, the file shares sometimes do not come online and the MSCS Cluster Administrator shows a status of "File share resource failed" (Incident 84803). This issue has been resolved by the hotfix so that the file share resources always show an online status after a cluster node is started or resources are moved.

4. When an MSCS Volume Manager Disk Group resource is moved from one cluster node to another, multiple SCSI bus resets may be performed on shared buses on the target machine. This slows the move operation (Incident 78221).
Also, for some hardware configurations, there is a possibility that a SCSI bus reset may be sent to the wrong bus, with the result that the disk group resource on the target node fails to come online.
This hotfix limits SCSI bus resets to one reset per bus per move group operation and also ensures that bus resets are sent to the correct bus. After applying the hotfix, a message will be sent to the system event log each time a SCSI bus reset is performed.

5. At shutdown, a system running MSCS and Volume Manager 2.7 will occasionally hang and not shut down in the expected manner (Incident 88473). This issue has been resolved by the hotfix.

6. When a mirrored volume is in the process of being brought to a healthy state after a disk failure, data corruption and possibly a blue screen may occur if an attempt is made to read data from a preferred mirror while that mirror is resynching. (Incident 87557). The hot fix prevents data from being read from the failed mirror during resynching.

7. Systems running Volume Manager 2.7 report error messages in the event log when using the Terminal Service Client and running Perfmon to add performance counters (Incident 88678).
This issue has been resolved by the hotfix and the user will not receive error messages when connected using the Terminal Service client.

8. The system will blue screen if an attempt is made to import a dynamic disk group that contains a disk with the same disk signature as the boot disk already being used (Incident 87528).

After the hotfix is applied, Volume Manager will not allow a disk group to be imported if one of the disks in the dynamic disk group has the same disk signature as the boot disk; instead an error message will appear explaining why the import procedure failed.

B. Past Issues from Previous HotFixes

Volume Manager 2.7 Hotfix 3

1. After rebooting, volumes containing Microsoft SQL Server 2000 databases show a status of "Healthy, Failed." An alert message will appear in the Events tab (Incident 77780).
This situation occurs only on systems running Volume Manager 2.7 for Windows and Microsoft SQL Server 2000. This issue is resolved by the hotfix. Microsoft SQL Server 2000 volumes will now show a "Healthy" status in Volume Manager 2.7 after reboot.

2. In any volume with three or more mirrors, rebooting the system or importing the disk group while two or more of the disks containing the mirrors are removed from the system may cause the removed disks to show a status of "Offline, Foreign" when they are reinserted in the system (Incident 78582).

Normally, disks that are missing during boot/import operations are automatically resynchronized after they are reinserted. However, the disks will show a status of "Offline, Foreign" if a different disk is installed and used each time the boot or import operation is performed while the remaining disks of the mirrored volume are removed from the system.
In this case, the reinserted disks will show a status of "Offline, Foreign" by design. A fix has been implemented to allow the user to select the command "Revert disk to basic" so that the disk can be returned to basic status and reused.

3. The system will fail to boot from a mirror volume if the size of the original mirror volume was exactly equal to the entire size of the disk on which the new mirror was created (Incident 85963).
This issue has been resolved by the hotfix. After applying the hotfix, any mirror in a healthy mirrored system volume will boot the system as expected.

4. After a missing disk is removed from Volume Manager 2.7, the Event Log shows an incorrect alert message "reverted disk to basic" (Incident 79152).

The alert message has been corrected in the hotfix and now reads "Successfully removed a missing disk."

5. In certain situations, the Rescan command causes the "Upgrade to Cluster Disk" menu option to appear for internal disks (Incident 70808).
This hotfix grays out this menu option to prevent the possibility of upgrading internal disks to a cluster disk group.

6. In certain situations, when a hard disk that is part of a dynamic disk group is pulled out from a server, a blue screen with the bugcheck KMODE_EXCEPTION_NOT_HANDLED may result (Incident 83809).
This issue has been resolved by the hotfix.

7. Improvements were made to the vxio driver to ensure that shutdown and flush commands are sent to lower drivers (Incident 69877).
These improvements minimize any potential loss of cached data in RAID controllers during shutdown.

8. This hotfix allows integration between SANPoint Control 3.5 and Volume Manager 2.7 (Incidents 85271, 86223).


Volume Manager 2.7 Hotfix 2

Bugcheck with stop code "0X50 PAGE_FAULT_IN_NONPAGED_AREA" (Incident 85793).

Occasionally, a system running Volume Manager 2.7 under heavy I/O may fail to load a driver correctly and result in a bugcheck with stop code 0x50, "PAGE_FAULT_IN_NONPAGED_AREA."
This problem has been resolved with the hotfix.


Volume Manager 2.7 Hotfix 1

1. Optimized the Rescan function to improve performance when hot plugging disks (Incident 71196).

Improved the rescanning process so that it rescans only for the disks that were removed or replaced rather than all the disks on the system.

2. The hot fix resolves a number of problems that caused HCT MSCS Validate 2 Node tests to fail when Volume Manager cluster disk group resources were in use. These problems might also affect normal operations or occur during certain failure scenarios. The following issues were resolved:

- Improvements were made to the processes for locking and dismounting volumes on cluster disk groups during an offline operation (Incident 53844).

NOTE: For cluster disk groups, Volume Manager makes several attempts to lock and dismount volumes before deporting the group, but eventually will force deport a group with active I/O. The cluster administrator should structure cluster group dependencies to ensure that applications that depend on a Volume Manager disk group resource are offlined before the cluster disk group is offlined.

- Performance for offlining cluster disk groups was improved (Incidents 73365, 73370).

- When the MSCS cluster service was stopped on a node with an online Volume Manager Disk Group resource, the resource could not be brought online again on that node until the node had been rebooted (Incident 73368).

This problem has been resolved with the hotfix.

3. In a system with MSCS and DMP installed, disabling both paths to a disk that is a part of a cluster resource causes the Cluster Administrator to freeze and not failover to the other node (Incident 72343).
This problem has been resolved with the hotfix.

4. After Volume Manager and DMP have been installed, adding or removing a SCSI adapter card may result in a blue screen with the error "Bugcheck 0x1e" (Incident 71033).
This problem has been resolved with the hotfix.

5. In a system that has Volume Manager with DMP and a Compaq Smart Array controller installed, the message "Driver Power_State_Failure STOP" may occur when there is a second shutdown after the PNP driver discovers new hardware (Incident 72454).
This problem has been resolved with the hotfix.



Attachments

VM2K27HF04JPN_245022.exe (5.7 MBytes)

Supplemental Materials

SourceiTools
Value76870
DescriptionWhen a resource is offlined or terminated, error 4 occurs.

SourceiTools
Value76871
DescriptionDuring rapid failover of MSCS resources on a multiprocessor server, bugcheck 0xOa can occur.

SourceiTools
Value76877
DescriptionWhen multiple resources are offlined and onlined simultaneously, the Volume Manager application can hang.

SourceiTools
Value76878
DescriptionWhen resources are being failed over continuously, the Volume Manager application can AV (Access Violation).

SourceiTools
Value76879
DescriptionImproved the speed of the failover process.

SourceiTools
Value84803
DescriptionWhen an MSCS cluster node is started or cluster resources are moved, the file shares do not come online...

SourceiTools
Value78221
DescriptionWhen an MSCS Volume Manager Disk Group Resource is moved from one cluster node to another...

SourceiTools
Value88473
DescriptionAt shutdown, a system running MSCS and Volume Manager 2.7 will occassionally hang and not shut down in the expected manner.

SourceiTools
Value87557
DescriptionWhen a mirrored volume is in the process of being brought to a healthy state after a disk failure...

SourceiTools
Value88678
DescriptionSystems running Volume Manager 2.7 report error messages in the event log when using the Terminal Service Client...

SourceiTools
Value87528
DescriptionThe system will blue screen if an attempt is made to import a dynamic disk group that contains a disk...

SourceiTools
Value77780
DescriptionAfter rebooting, volumes containing Microsoft SQL Server 2000 databases show a status of "Healthy, Failed."

SourceiTools
Value78582
DescriptionIn any volume with three or more mirrors, rebooting the system or importing the disk group while two or more of...

SourceiTools
Value85963
DescriptionThe system will fail to boot from a mirror volume if the size of the original mirror volume was exactly equal to the entire...

SourceiTools
Value79152
DescriptionAfter a missing disk is removed from Volume Manager 2.7, the Event Log shows an incorrect alert message...

SourceiTools
Value70808
DescriptionIn certain situations, the Rescan command causes the "Upgrade to Cluster Disk" menu option to appear for...

SourceiTools
Value83809
DescriptionIn certain situations, when a hard disk that is part of a dynamic disk group is pulled out from a server, a blue screen...

SourceiTools
Value69877
DescriptionImprovements were made to the vxio driver to ensure that shutdown and flush commands are sent to lower drivers.

SourceiTools
Value85271, 86223
DescriptionThis hotfix allows integration between SANPoint Control 3.5 and Volume Manager 2.7.

SourceiTools
Value85793
DescriptionOccasionally, a system running Volume Manager 2.7 under heavy I/O may fail to load a driver correctly...

SourceiTools
Value71196
DescriptionOptimized the Rescan function to improve performance when hot plugging disks.

SourceiTools
Value53844
DescriptionImprovements were made to the processes for locking and dismounting volumes on cluster disk groups during an offline operation.

SourceiTools
Value73365, 73370
DescriptionPerformance for offlining cluster disk groups was improved.

SourceiTools
Value73368
DescriptionWhen the MSCS cluster service was stopped on a node with an online Volume Manager Disk Group Resource...

SourceiTools
Value72343
DescriptionIn a system with MSCS and DMP installed, disabling both paths to a disk that is part of a cluster resource...

SourceiTools
Value71033
DescriptionAfter Volume Manager and DMP have been installed, adding or removing a SCSI adapter card may result...

SourceiTools
Value72454
DescriptionIn a system that has Volume Manager with DMP and a Compaq Smart Array controller installed, the message...


Legacy ID



245022


Article URL http://www.symantec.com/docs/TECH18592


Terms of use for this information are found in Legal Notices