Technically, different nodes of the cluster may be at different OS patch levels. However, this is not considered Best Practice. Previous to Veritas Cluster Server (VCS) version 5.1 differing OS versions were not supported.
- Best Practice is to have each node of the cluster to be exactly alike to ensure that all applications and services will function correctly in each host environment.
- This is true in both RAC and failover environments.
- RAC: It will be necessary that the application function correctly in the different OS environments so all nodes of the cluster will perform simultaneously.
- Failover: In the event of a failover, if the affected application or function (Service Group) fails to run in the different OS environment of the other node (commonly called the failover node), the VCS failover will be unable to bring the Service Group online on the failover node. If the affected application malfunctions (runs but fails to function properly) on the failover node, it will appear that VCS has malfunctioned due to the application or service malfunction in the different OS environment of the failover node.
This is generically true in all envirmonents where VCS is supported.