This example upgrade uses four nodes with four service groups.

• The cluster is split into two subclusters. The first subcluster consists of node01 and node02 which is considered for upgrading first. The second subcluster consists of node03 and node04 which will be upgraded second.

• Switch all the failover service groups from the first subcluster to the second subcluster. Take all the parallel service groups offline on the first subcluster. Freeze the service groups.

• Stop VCS and all its components including GAB and LLT.

• Unfreeze the service groups on the second subcluster.

• Upgrade the operating system, if required.

• Upgrade VCS on the first subcluster. After the upgrade is complete, not start any of the cluster component, including VCS, GAB, and LLT.

  Note:

  Your downtime starts from here.

• Now, you can proceed with upgrading the second subcluster.

• Take all the service groups offline on the second subcluster. Stop VCS and its components.

• Reboot the nodes in the first subcluster. Unfreeze all the service groups and bring them online.

  Note:

  Your downtime ends here.

  Note:

  After the first subcluster restarts, make sure that the nodes on the second subcluster do not reboot accidently. If the second subcluster restarts and joins the first subcluster, the nodes in the second subcluster may panic.

• Upgrade VCS on the second subcluster. After the upgrade is complete, do not start VCS. Reboot the nodes in the second subcluster.

  The nodes in the second subcluster automatically rejoin the cluster. All the nodes in the cluster now run the upgraded version of VCS.