Very recently I had 1 of our 6 Hypver-v hosts cease to boot from the san.
So I removed it from the Hyper-V Cluster and reinstalled windows and brought it back up with the same name.
I rejoined it to the Hyper-V cluster. Everything went well enough until the next morning when I found that I could not put any of our 6 servers into Maintenance mode from the SCVMM 2012 console. In the console I get this message:
Warning (23756)
Changes have been made which should be verified by running placement again.
Recommended Action
Use the Refresh Preview action to run placement again.
I can live migrate single VM's around just fine and I can even clear out an entire host by doing a pause - drain roles from the failover cluster management tool.
Once a host is empited out and not running any VM's then it can go into maintenance mode but what good is that?
I have gone down the rabit hole of Trace logging and this is what I am staring at:
E2C.0005::10/22-14:56:01.842#19Task.cs(259): Task failed error VMConfigurationPlacementOutOfDate (23756) with exception Microsoft.VirtualManager.Utils.CarmineException: Changes have been made which should be verified by running placement again.
00023123 189.36448669 [3628] Use the Refresh Preview action to run placement again. ---> Microsoft.VirtualManager.Engine.ImageLibrary.VMChildObjectNotFoundException: VMM cannot find VirtualHardDisk object e611e3a9-e606-4b67-a2e9-dd88924ad7ff.
I have gone the router of backing up my Hyper-V database and running a couple of different sql scripts that are intended to remove orphaned objects. This has not helped.
I can not find this VHD object anywhere I have looked in the DB.
Does anyone have any ideas on how I may fix this issue. I am lucky in that all VM's are up and running but I am concerned for what this may do to our High Availability.