Veritas InfoScale™ 7.4.1 Installation and Upgrade Guide - Windows

To upgrade SQL Server to a service pack, perform the following steps:

1. Stop the replication between the primary and the secondary site.

   If using Volume Replicator for replication, from the VEA Console, right-click the Secondary RVG and select Stop Replication from the menu that appears.

2. On Node A where the SQL Server service group is online, take the SQLServer, MSOlap, and SQLServer-Agent resources offline.

   Run the following command from the command prompt:

   hares -offline [-parentprop] resource -sys system

   Here, resource is the name of the SQL resource and system is the name of node where the SQL Server service group is online.

3. From Services.msc, ensure that all the SQL services and the SQL services for which VCS resources are configured are stopped.
4. Perform the following steps on the SQL Server service group on Node A (active node):

   • Bring the RegRep resource offline.

     Type the following on the command prompt:

     hares -offline [-parentprop] resource -sys system

   • Disable the RegRep resource.

     Type the following on the command prompt:

     hares -modify resource_name Enabled 0

   • Except the storage resources (MountV and VMDg), take all the resources offline.
     

   • Take a backup of the SQL Server directories from the shared storage.
     Then freeze the service group.

     Type the following on the command prompt:

     hagrp -freeze service_group [-persistent]

5. Install the Microsoft SQL Server Service Pack on Node A.
6. If a FILESTREAM resource is configured in the SQL Server service group, verify if a FILESTREAM share exists on the node and then delete it.

   Run the following commands from the command prompt:

   • net share

     This command lists all the shares on the node.

   • net share share_name /delete

     Here, share_name is the name of the FILESTREAM share.

7. Unfreeze the SQL Server service group.

   Type the following on the command prompt:

   hagrp -unfreeze service_group [-persistent]

8. Fail over the service group to Node B and perform the following steps on Node B, in the given order:

   • Except the storage resources (MountV and VMDg), take all the resources offline.

   • Open the Services window, and ensure that all the SQL Server services and the ones for which VCS resources are configured are stopped.

   • Freeze the service group.

9. Perform the following steps:

   • Rename the SQL Server folders on the shared storage and copy the backed up SQL Server directories to the shared storage.

   • Rename the SQL Server replication directory that is present under Registry Replication folder in the shared storage.

   The SQL Server data files available on the shared storage are upgraded during the SQL upgrade on Node A. Before you begin to upgrade SQL on Node B, you must rename the folders containing the upgraded SQL data files and restore the initially backed up SQL Server directories. If you do not restore the initially backed up SQL Server directories, then the SQL upgrade on Node B may fail indicating that the SQL Server data files are already upgraded.

10. Install the Microsoft SQL Server Service Pack on Node B.
11. If a FILESTREAM resource is configured in the SQL Server service group, verify if a FILESTREAM share exists on the node and then delete it.

    Run the following commands from the command prompt:

    • net share

    • net share share_name /delete

      Here, share_name is the name of the FILESTREAM share.

12. Unfreeze the service group on Node B and enable the RegRep resource.

    Run the following commands from the command prompt:

    • hagrp -unfreeze service_group [-persistent]

    • hares -modify resource_name Enabled 1

13. Bring the service group online on Node B.
14. Start the replication between the primary and the secondary site.
15. Switch the service group to a node on the DR site (Node C).

    Type the following on the command prompt:

    hagrp -switch service_group -site site_name

16. Stop the replication between the primary and the secondary site again.
17. Perform the following steps on Node C, in the given order:

    • Except the storage resources (MountV and VMDg), take all the resources offline.

    • Disable the RegRep resource and freeze the service group.

    • Rename the SQL folders on the shared storage and copy the backed up directories to the shared storage.

    • Install the Microsoft SQL Server Service Pack on Node C.

    • If a FILESTREAM resource is configured in the SQL Server service group, verify if a FILESTREAM share exists on the node and then delete it using the following commands:

      net share

      net share share_name /delete

    • Unfreeze the service group and enable the RegRep resource.

18. Start replication between the primary and secondary site.
19. Switch the service group back to Node B (last upgraded node) on the primary site.

    Note:

    You must bring the SQL service group online on Node B first. This is because the replication service group is online on Node B. You can then switch the SQL service group on any node on the primary site.

To upgrade SQL Server to a service pack

1. Freeze the SQL Server service group on Node A using the VCS Cluster Manager (Java Console).

   On the Service Groups tab, right-click the service group and then click Freeze > Persistent. Save the configuration and leave the group frozen until all the nodes are updated.

   Alternatively, run the following command on the command prompt:

   hagrp -freeze service_group [-persistent]

2. Back up the registry of the SQL Server database instance on the passive nodes, by following these steps sequentially:

   • Access the registry by using your preferred method to execute regedit.exe.

   • Locate the SQL Server instance registry key:

     HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft SQL Server\MSSQL11.instance

   • Right-click on the registry key and click Export.

   • Save the registry key.

   Note:

   While upgrading SQL Server 2014 or SQL Server 2016 to a service pack, the MSSQL11 value in the registry key changes to MSSQL12 or MSSQL13 respectively.

3. Stop the replication between the primary and the secondary sites.

   If VVR is used for replication, from the VEA Console right-click the Secondary RVG and select Stop Replication from the context menu.

4. If you find any registry keys from the following list on nodes B and C, update them to point to the root directory. Note that instance represents the SQL Server instance that you are trying to upgrade.

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\BackupDirectory

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\DefaultLog

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\Parameters\SQLArg0

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\Parameters\SQLArg1

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\Parameters\SQLArg2

   • HKLM\SOFTWARE\Microsoft\Microsoft SQL Server\MSAS11. instance\CPE\ErrorDumpDir

   • HKLM\SOFTWARE\Microsoft\Microsoft SQL Server\MSAS11. instance\Setup\DataDir

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\DefaultData

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\SQLServerAgent\ErrorLogFile

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\SQLServerAgent\WorkingDirectory

   • HKLM\Software\Microsoft\Microsoft SQL Server\MSSQL11.instance\Setup\SQLDataRoot

   For example, if the key HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft SQL Server\MSSQL11.instance\MSSQLServer\BackupDirectory contains the value F:\Test\MSSQL11.MSSQLSERVER\MSSQL\Backup, change this value to point to the root directory C:\Program Files\Microsoft SQL Server.

   Therefore, updated value becomes C:\Program Files\Microsoft SQL Server\MSSQL11.MSSQLSERVER\MSSQL\Backup.

5. Complete the SQL Server service pack upgrade on nodes B and C, and reboot the server.

   Do not fail over until the service pack installation on Node A is complete.

6. On Node A, run the command hastop -local -force to stop the VCS services.

   Doing so stops VCS, but leaves the storage and the SQL Server online. It also prevents VCS from taking any action on the resources when the installer restarts the service.

7. Complete the SQL Server service pack upgrade on Node A, and verify that the services are back in the Started state.
8. Run command hastart to start VCS services on Node A.
9. Start the replication between the primary and secondary sites.
10. Unfreeze the service group and test failover to nodes B and C.

    Run the following command on the command prompt:

    hagrp -unfreeze service_group [-persistent]