SPS for Linux may be upgraded to future releases while preserving existing resource hierarchies. Review this information carefully to ensure that you minimize application downtime.
- While upgrading a cluster, switch all applications away from the server to be upgraded now. Do this manually or by setting the LifeKeeper shutdown strategy to “Switchover” which causes the applications to be switched when LifeKeeper is stopped or the server is shut down.
- If necessary, upgrade the Linux operating system before upgrading SPS. It is recommended that you unextend all resources from a server that is to be upgraded prior to performing the operating system upgrade.
- Upgrade LifeKeeper by referring to How to Use Setup Scripts.
- Switch all applications back to the upgraded server.
- Repeat this procedure for each server in the SPS cluster to be upgraded.
When upgrading the OS, make sure the currently installed version of LifeKeeper supports the upgraded version of the OS. If it is not supported, LifeKeeper will need to be upgraded as well provided a version of LifeKeeper has been released that supports the new OS version. If no version of LifeKeeper has been released that supports the new OS version you may not be able to upgrade the OS. Refer to the Supported Operating Systems.
Before upgrading the OS, it is recommended that the LifeKeeper configuration be backed up via the lkbackup command.
- When upgrading the cluster, all the resource hierarchies and thus the applications they protect, must be switched from the server to be upgraded to a standby node in the cluster. This can be done manually, or, by setting the LifeKeeper Shutdown Strategy to “Switchover”. By setting the Shutdown Strategy to “Switchover”, the resource hierarchies are switched over to a standby node when LifeKeeper stops or the servers are shutdown.
- Stop LifeKeeper.
- Upgrade and reboot the OS.
- Upgrade LifeKeeper when required. If you do not upgrade, run setup again and update the settings corresponding to the new OS.
- Start up LifeKeeper.
- Switch all the resource hierarchies to the upgraded server.
- Execute these steps for all the nodes in the SPS cluster.