(imported topic written by SystemAdmin)
We’d like to see some support for intelligently patching MS Clusters. Exchange and SQL are the two big ones. What we would like to see is are two different ways to patch cluster hosts.
Type 1 - Patching of passive nodes with all non-clustered patches
This would apply all the patches to the passive nodes - Normal windows patches and application patches that do not require the app to be running (Some exchange and sql patches are like this).
Type 2 - Patching the active nodes with cluster application patches
This would apply all the patches which were dependent on the clustered app being running.
So in an automated fashion, type 1 patches would execute first on all passive nodes. Passive nodes would report back success. Then type 2 patches would excecute on all active nodes (one node at a time per cluster). The active / passive nodes would then failover. Then the previously active / now passive nodes would run type 1 patches. This would be followed by the now active / previously passive nodes would then run the type 2 patches.
I hope that wasn’t tooooo confusing