Remove Snapshot stuck at 95% after Veeam Backup - Help!

[rpmahony]

Normally, I'd just be patient.  But when I didn't see my "Backup Completed" email from Veeam from Saturday's full… by Monday evening…  I hopped on the backup server to take a look.  The backup status of my Exchange 2010 server was sitting at 100% completed, waiting for the snapshot to be removed… for 56 hours!  Normally, this machine backs up in about 8 hours.  I've been checking file sizes of the deltas… and they're growing instead of shrinking.  I've tried googling several iterations of "remove snapshot stuck at 95%" but frankly, I'm scared to try anything without first consulting the community.  Has anyone seen this issue with ESX 3.5 and have experience with getting things cleared up without killing the guest OS?  I'm even scared to try to reboot the VM to clear potential VSS issues that might be causing the issue (though I don't see any VSS event log entries.

Briefly in the early morning hours, I did see a few messages in the VI recent tasks that stated "the parameter was not correct".

Where do I start?!  Help!

Rusty

[a_p_]

>>> waiting for the snapshot to be removed…

Where do you see a stuck task? Only in the backup application, or also on the ESXi host? Maybe it's just a communication issue where the snapshot removal failed, and the backup application is still waiting for feedback!? If the task doesn't show up on the host (i.e. connected to the host directly) anymore, then it most likely doesn't run anymore, and you should be able to manually "Delete" the snapshot from the Snapshot Manager. If no snapshot shows up in the Snapshot Manager, create a new one and then run "Delete All".

One thing you should check is whether there's free disk space on the datastore, because the second/new snapshot will grow until deleting the snapshots has finished.

André

[rpmahony]

Remove Snapshot is showing up in the VI Client, attached directly to the host.  The datastore is showing 299GB free space, so I'm ok for now on that end… I think.  The growth rate isn't super big at present but I'm getting a little nervous that it's not going the other direction.

[ch1ta]

In the process of creating and deleting snapshot Veeam plays only "requester" role, which means it just issues a certain API calls to vCenter/ESX(i) hosts. So, I would exclude Veeam from this equation, and try to understand why the underlying host can't cope with snapshot removal operation. If possible, I would open a ticket with VMware.

Cheers.