Backup Copy Job. VM Retention.

[emilec]

Can I throw something else in here as well. Case: 00561220

There seems to be some confusion as to the expected behaviour of the VM retention tickbox with copy jobs. My understanding for normal backup jobs is that when I delete the VM from the backup job it will get cycled out of the backups once the retention period is passed. I had a copy job that processed some machines, failed due to a connection reset and then didn't run for 2 days (no new local backups on Weekends). My retention was set to 3 days. When the copy job next ran it deleted all the machines that had restore points older than 3 days due to the vm retention. At the time when the copy job started the machines it deleted were still included in the copy job and there was new data on the local source repository. So it would appear to me that it was looking at what was on the remote repository rather than what was included in the copy job.

[veremin]

I'm a little bit confused. Are you talking here about deleted VM retention period ("remove deleted VMs")? Thanks.

[emilec]

That's the one. Referenced here: http://helpcenter.veeam.com/backup/70/v ... tings.html

[veremin]

How long was the backup copy chain? Can you confirm that there were successful increments for those VMs (that got deleted then) in backup copy chain, starting from full backup? Thanks.

[emilec]

This was an initial seed from a full job and the first copy job that ran. All the VMs were in the initial seed, those that managed to complete in the first copy run remained, the rest were deleted.

[foggy]

So basically the only successful restore point in the chain for those VMs that were finally deleted was the full one, right? And then there were no successful job runs during the specified retention period (3 days)?

[emilec]

That is correct.

I would however still not expect them to have been deleted as they were still included in the job.

[foggy]

The logic behind this setting is not tied to the list of VMs in the job, actually. Retention is also applied in the case, where VM is removed from infrastructure, for example, or has migrated to a different host/datastore than the one added to the job (if backing up by host/datastore). The fact is that the VM data is removed from the backup if it was not successfully backed up (for any reason) at least during the specified number of days and there are no successful incremental restore points containing it.

[emilec]

Ok, so you're saying this is expected behaviour?

[foggy]

Yes. I would recommend to increase the number of days to keep deleted VMs to avoid such situations.

[emilec]

Ok thanks. Then only other thing then is that this doesn't come across clearly in the documentation I referenced. So possibly this can be expanded a bit?

[foggy]

Yep, the description could be expanded. I will pass this to the documentation team. Thanks!

[emilec]

Thanks for all the help!