[Hyper-V] SQL tables HvVolumeSnapshots and Host not cleaned

[RGijsen]

Case 00882079. I've been asked to report here to notify R&D. In short; after removing three hosts from our 5-node Hyper-V cluster and after that running one of those nodes as stand-alone Hyper-V host (keeping the same hostname), the HvVolumeSnapshots and Host table weren't cleaned up. The snapshot table had 127 references to snapshots on the host that was running as standalone now. That made backups extremely slow between phases. The logfiles reflected Veeam was trying to locate the host and snapshots removed from cluster. The automatic 'refresh topology' that runs every ... 4? hours did remove the 2 hosts and took one out of the cluster. From GUI perspective it was all fine. It just didn't cleanup the snapshots table as there were all snapshots there of machines running on the cluster, that are unable to run on the standalone host.

So basically the 'refresh topology' did not work too well in this case. There could be some more checks / housekeeping.

[foggy]

Actually, those records should not stay in the database but rather cleaned up on the job completion, so the fact they are still there should be investigated (not automatic "refresh topology" issues). I've passed this to QA for review, thanks for the heads up.