Updating a SLES11 Cluster to the latest Service Pack

An existing SLES 11 HAE Cluster should be updated to SLES11 SP1 HAE or an existing SLES 11 SP1 HAE Cluster should be updated to SLES11 SP2 HAE with the least downtime possible.

The technical process of updating is described in TID 7005410 - How to upgrade to SLES/SLED 11 SP1

Further consideration is the behaviour of the cluster itself. As the SLES11 HAE cluster stack is for High Availability the failure of one cluster node won't affect the other machines. To use this the following process should be followed on each node,  one after the other:

1. Disable the cluster service in the runlevel on a node

   insserv -r openais

to ensure, that if anything goes wrong on this node during the update no unhealthy node tries to rejoin the cluster

2. Move all cluster resources away from this node by putting this node into standby, for example via commandline:

   crm node standby NODE_NAME

3. Stop the cluster service on a node

   rcopenais stop

If the stop fails the machine can also be stopped or rebooted at this stage as it won't rejoin the cluster.

4. Update the Software as described in the TID linked above.

5. Reboot the node.

6. Start the cluster service on this node.

   rcopenais start

7. Allow cluster resources back onto this node by making it online in the cluster again, for example via commandline:

   crm node online NODE_NAME

8. Add cluster service to runlevel:

   insserv openais

Repeat process as outlined above for each and every node consecutively.