Found situation where Veeam deleted my backups

[zoltank]

So over the weekend I found a situation where Veeam deleted most of my backups. I have Veeam installed on a server two buildings up from where the server room is, in a remote office there. It does an incremental backups on Sunday, and then creates a synthetic full. It's set to retain 1 restore point (it's for DR only).

The problem occurred when one of the VMs in one of its backup job ran low on disk space and wasn't able to successfully create a VSS shadow copy, as a result the backup of that VM failed, and then failed again on the retry. I then fixed the problem and told the job to manually retry. As is normal it retired just the one failed VM and was able to successfully back it up. However, after the successful backup the backup only contained the one failed VM and none of the other VMs. Running an Active Full fixed the situation.

It looks like the other VMs were removed from the backup even before I fixed the situation and backed up up the VM successfully. My guess is the retry counted as a restore point so Veeam flushed the restore point with the good backups of the other VVMs.

I opened a support case (00254092) to let Veeam know.

[zoltank]

Veeam says this is a known issue which will be resolved in v.7.

In the meantime increasing your restore points will prevent this from happening.

[foggy]

Zoltan, could you please clarify regarding retention settings for this job: how many restore points it is configured to keep and what is the Deleted VMs retention period for it?

[zoltank]

It was originally configured to keep only 1 restore point.

The deleted VM retention period is set to 14 days.

[foggy]

I've tried to dig this deeper and it seems that the particular fix the support guy is talking about will not help in your specific case. It will address cases where the entire job failed (creating the empty restore point) but the oldest restore point was still deleted according to retention policy. The new (fixed) logic deals with those empty points and does not apply retention in that case. In your case, however, there were some successful VMs in the new restore point and retention will still be applied. So increasing the number of restore points to be kept seems to be the only way to avoid that.

[zoltank]

I had already increased my restore points before even opening the case.

I'd recommend making it every clear in the documentation this is a possible scenario when using a single restore point. Additionally, could it be changed so the software sees the successful retry as part of the existing restore point instead of a new one?

[foggy]

R&D will try to address this use case in future.