CSV ran out of space during snapshot merge – now have too many AVHD files and a merge-pending state

While doing upgrade work on a VM running on a Windows Server 2016 cluster, multiple snapshots were created to provide a rollback if the upgrades had issues (they did, multiple times, so this was very helpful to have).  Once upgrade was completed and
working correctly, I deleted the newest snapshot.  Another tech had, in the meantime, created a new VM that used most of the remaining space on the CSV.  So, my VM moved to Paused-Critical state when the merge used up the remaining space.  I
immediately storage migrated other, still running VMs from the out-of-space CSV to another one.  I then resumed the VM that was in the Paused-Critical state.

The VM appears to be fine and operating correctly.  I can’t find any signs of data loss.  The snapshot I deleted no longer shows in Hyper-V manager.  Inspecting the disks show an intact differencing disk chain that has 3 links in it (Base
VHD, AVHD for the remaining snapshot, and AVHD for "Now").  This is what I would expect.

My concern is that there are 2 additional AVHD files on the CSV that are not chained into running configuration.  And an examination of the VM XML config file (version 2) shows that disk_merge_pending is true.

My suspicion is that there when I shut down the VM, it will try to finish the merge that was aborted when space ran out on the CSV.  I am not sure what affect this will have on the state of the data, since and inspection of the current chain no longer
includes these AVHD files.  Or, it may be that everything completed once space became available, and the flag simply didn’t get cleared or the AVHD files deleted.

Can anyone offer any insight into what I can expect when I shut this machine down?  Should my data remain intact?

Obviously I will take a backup immediately before shutting down, but I don’t want to have to do a restore if I don’t have to.

Thanks in advance for any help or insights.

Tim Reid

Read full post . . . .

Go Que Newsroom