Upgrading the cluster subsystem

You can upgrade the cluster subsystem in two ways:

• Perform a major upgrade, which involves cluster down time, then shutting down and restarting all servers. In this case, you must:

  1. Run sp_companion suspend, as described in Using Sybase Failover in a High Availability System.

  2. Take the resource groups of primary and secondary companions on both nodes offline. Make sure this shuts down both the companion server and corresponding resource groups, and that they are not automatically brought online until the cluster system upgrade is complete.

  3. Upgrade the cluster subsystem following the instructions from the cluster system vendor.You may find options to migrate the current resource groups to the new cluster version. If such an option is not available (or if the resource groups get deleted or corrupted), re-create the resource groups and configure them appropriately after you have upgraded the cluster system.

  4. Bring the resource groups online. This should bring primary and secondary companions online on their respective nodes.

  5. Run sp_companion resume as described in Using Sybase Failover in a High Availability System.

• Perform a minor upgrade, which does not involve a cluster downtime. In this case, each node is failed over to another node and upgraded one at a time. Assuming 1 is the primary companion on node N1, and 2 is secondary companion running on node N2:

  1. Upgrade the primary companion:

     1. Fail over 1 from N1 to N2. You can achieve this by relocating the primary resource group to N2 or by shutting down 1.

     2. Upgrade the cluster subsystem on N1 following the upgrade instructions provided by the vendor.

     3. Fail back 1 from N2 to N1. See the appropriate cluster chapter in Using Sybase Failover in a High Availability System for more details on Adaptive Server failback.

  2. Upgrade the secondary companion.

     • If you are using an asymmetric configuration:

       1. Take the secondary resource group offline and verify 2 is shut down. 2 is not available during this upgrade.

       2. Upgrade the cluster subsystem on N2 following the upgrade instructions from the vendor.

       3. Start 2 by bringing the secondary resource group online to N2.

     • If you are using a symmetric configuration, follow the steps in “upgrade the primary companion,” above, for 2.