To start the rolling upgrade - phase 1

1. Perform the required steps to save any data that you wish to preserve. For example, take back-ups of configuration files.
2. Complete the preparatory steps on the first sub-cluster.

   See Preparing to perform a rolling upgrade to SF Oracle RAC 6.0.1.

3. If you are upgrading to Solaris 10 Update 10, apply the following Oracle (Solaris) patches. For instructions, see Oracle documentation.

   For SPARC: 144524-02

   For x86: 144525-02

   Complete updates to the operating system, if required.

   Note:

   Make sure that the operating system update you apply is supported by the existing version of SF Oracle RAC. If the existing version of SF Oracle RAC does not support the operating system update, first upgrade SF Oracle RAC to a version that supports the operating system update. For example, if you plan to apply Solaris 10 Update 9 to SF Oracle RAC version 5.1 running on Solaris 10 Update 6, you need to perform the following steps before proceeding with the steps in this procedure: First, upgrade SF Oracle RAC to version 5.1 SP1RP2. For instructions, see the product documentation of that release. Next, update the operating system.

   For instructions, see the operating system documentation.

   The nodes are restarted after the operating system update.

   Restart the nodes again manually. Failing to perform this additional reboot prevents the upgrade from proceeding further.

   # shutdown -g0 -y -i6

4. Start the Web-based installer, if it is not already running.

   See Starting the Veritas Web-based installer.

5. In the Task pull-down menu, select Rolling Upgrade.

   Click the Next button to proceed.

6. Enter the name of any one system in the cluster on which you want to perform a rolling upgrade.The installer identifies the cluster information of the system and displays the information.

   Click Yes to confirm the cluster information. The installer now displays the nodes in the cluster that will be upgraded during phase 1 of the upgrade.

7. Review the systems that the installer has chosen for phase 1 of the rolling upgrade. These systems are chosen to minimize downtime during the upgrade.

   Click Yes to proceed.

   The installer validates systems. If it throws an error, address the error and return to the installer.

8. Review the End User License Agreement (EULA). To continue, select Yes, I agree and click Next.
9. If the upgrade is from the 5.1 SP1 release or later and the boot disk is encapsulated and mirrored, you can create a backup boot disk.

   If you choose to create a backup boot disk, type y. Provide a backup name for the boot disk group or accept the default name. The installer then creates a backup copy of the boot disk group.

   If you choose to create a backup boot disk, provide the name of the backup disk group or accept the default name.

   Additionally, select the following options:

   • Split mirrors on all the systems

   • Use the same disk group name on all the mirrored systems

   Select the systems where you want to create a backup of the boot disk.

   Click Next. The installer then creates a backup copy of the boot disk group.

   See Re-joining the backup boot disk group into the current disk group.

   See Reverting to the backup boot disk group after an unsuccessful upgrade.

   The installer lists the packages to upgrade on the selected node or nodes.

10. If you have online failover service groups, the installer prompts you to choose to switch these service groups either manually or automatically. Choose any option and follow the steps to switch all the failover service groups to the other subcluster.
11. The installer stops all processes. Click Next to proceed.

    The installer removes old software and upgrades the software on the systems that you selected.

12. If you want to enable volume or file replication or global cluster capabilities, select from the following options:

    • Veritas Volume Replicator

    • Veritas File Replicator

    • Global Cluster Option

    Click Register to register the software. Click the Next button.
13. When prompted by the installer, reboot the nodes on the first half of the cluster.

14. Perform the following steps if the subcluster has non-global zones.

    • Take the zone service groups in the subcluster offline:

      # hagrp -offline zone_group -sys sys_name

    • From the global zone, run the following command for each non-global zone in the subcluster:

      # zoneadm -z zone_name attach -u

    • Uninstall and install version 6.0.1 of the VRTSodm package on each global zone of the subcluster:

      # pkgrm VRTSodm
      # pkgadd -d vrtsodmpkg_path

    • Boot each non-global zone:

      # zoneadm -z zone_name boot

    • Log in to each non-global zone of the subcluster and run the following command:

      # svcadm enable -r vxodm

    • Halt each non-global zone:

      # zoneadm -z zone_name halt

    • Bring the zone service group on the subcluster online:

      # hagrp -online zone_group -sys sys_name

15. Manually mount the VxFS and CFS file systems that are not managed by VCS.
16. Relink the SF Oracle RAC libraries with Oracle by choosing the option Relink Oracle Database Binary from the program menu.

    Note:

    If the subcluster has non-global zones, you need to relink only the ODM library.

    See Relinking Oracle RAC libraries with the SF Oracle RAC libraries.

17. If the boot disk is encapsulated, the installer strongly recommends a reboot of the nodes. Reboot the nodes as prompted by the installer.

    Note:

    Before you reboot the nodes, ensure that the boot device is set to the disk containing the upgraded version of the product. # eeprom

18. Bring the Oracle database service group online.

    • If VCS manages the Oracle database:

      # hagrp -online oracle_group -sys node_name

    • If VCS does not manage the Oracle database:

      # srvctl start database -d db_name

19. Start all applications that are not managed by VCS. Use native application commands to start the applications.
20. When prompted, perform step 2 through step 19 on the nodes that you have not yet upgraded.

21. • If VCS manages the Oracle database, reset the AutoStart value to 1 to enable VCS to bring the database service group online when VCS starts:

      # haconf -makerw
      # hagrp -modify oracle_group AutoStart 1
      # haconf -dump -makero

    • If VCS does not manage the Oracle database, change the management policy for the database to automatic:

      $ srvctl modify database -d db-name -y AUTOMATIC

22. Migrate the SFDB repository database.

    See Post upgrade tasks for migrating the SFDB repository database.

23. When prompted, start phase 2. Click Yes to continue with the rolling upgrade.

To upgrade the non-kernel components - phase 2

1. The installer detects the information of cluster and the state of rolling upgrade.

   The installer validates systems. Click Next. If it throws an error, address the error and return to the installer.

2. Review the End User License Agreement (EULA). To continue, select Yes, I agree and click Next.
3. The installer stops the HAD and CmdServer processes in phase 2 of the rolling upgrade process. Click Next to proceed.
4. The installer removes old software and upgrades the software on the systems that you selected. Review the output and click the Next button when prompted. Register the software and click Next to proceed. The installer starts all the relevant processes and brings all the service groups online.
5. If you have network connection to the Internet, the installer checks for updates.

   If updates are discovered, you can apply them now.

6. A prompt message appears to ask if the user would like to read the summary file. You can choose y if you want to read the install summary file.
7. Upgrade Oracle RAC, if required.

   Note:

   Oracle RAC 11g Release 1 Clusterware is not supported. Make sure that you install Oracle RAC 11g Release 2 Grid Infrastructure in order to use the Oracle RAC 11g Release 1 database. All database versions starting from Oracle 10g Release 2 and later are supported.

   For instructions, see the chapter Upgrading Oracle RAC in this document.