I was wondering how your back-out plans look like. Untill now I must admit I never have had a really decent back-out plan in place when installing cumulative updates. I did ran into some issues sometimes, but usually those were fixable in a small amount of time inside of the outage window. But I would not have a good solution when the shit really hits the fan.
Theoretically speaking, would this suffice:
– Restore pre-upgrade SQL back-ups of all databases
– Restore pre-upgrade VM images of front-ends and back-ends
I cannot find a reason why it wouldn’t, but since I haven’t tried….
We’re also running DocAve back-ups at the customer I’m at; but do those also revent SharePoint binaries? I assume they don’t.
The SharePoint Object Model for backup does not cover the SharePoint binaries themselves. You would need to restore the entire farm from a pre-upgrade backup in order to ‘back out’ of the specific CU. I would never recommend this — instead, make sure to do full test passes on a pre-production farm that is configured as close to the production farm as possible. You won’t weed out every issue, but you may catch business critical issues. Do keep in mind that snapshotting or taking live bacukps SharePoint VMs is not supported.
