Full Running Twice

[nunciate]

Anyone else seeing this issue with Veeam Tape Out?
I run synthetic full backups with reverse incrementals on Fridays. I have noticed this happen on random jobs.

On Friday all of my jobs ran. I can see the logs for this 1 particular job and it says on Friday the backup ran and did its transform.
I also run tape jobs and they are set to start when the backup job finishes. The backup job finished at around 1:15am that night.
The corresponding tape job shows it started at the same time the backup finished. I see the tape job ran as a full, finished around 4:45am and transferred 1.6Tb so that all looks good.

No new backup runs on this system until Monday night so last night I see the normal incremental backup ran. However when the tape job ran it started a new weekly backup and did a full backup to tape.

Obviously something is happening to cause a full tape out but not sure what that could be. The job has not changed since Friday. No servers were added to the job and when I look at the log from the backup no transform of the full occurred.

[veremin]

Hi, Alan,

I'm wondering how you scheduled a backup to tape job to run after a backup job. More specifically, what option did you use? "After this job" or "As new backup file appears"? Additionally, what is specified as a source of backup to tape job? Individual job, bunch of jobs or repository as a whole?

Thanks.

[nunciate]

My schedule tab on the tape job has After this job selected and the job for backup to disk is selected.

[veremin]

And what is selected as a source for backup to tape job? Individual VMs, several VMs or repository as a whole? Thanks.

[nunciate]

This job has multiple VMs selected

[veremin]

I must have confused the type of "source" that backup to tape and backup copy job has. What was meant is, of course, "individual Job, multiple Jobs, etc.".

The assumption is that if the several jobs are selected as a source for backup to tape job, and this jobs have different "full" backup schedule, it might happen that two full backups that were transferred by backup to tape job belonged to two different backup jobs.

If it's not the case, kindly, open a ticket with our support team and let them investigate this behavior directly, since it's rather hard to speculate about the possible causes without seeing logs and actual environment.

I run synthetic full backups with reverse incrementals on Fridays.

Just to be sure - by "reverse incrementals" you've meant "synthetic full with rollback transform", not reversed incremental backup mode,right?

Thanks.

[trustblomme]

Is your backup to disk retention lower than your backup to tape retention?

Steven

[nunciate]

I found the problem. This backup job actually does a backup of one of my Veeam proxy servers. I like to back that up as I keep some data on a share on there. What happened was that the backup proxy had done a hotadd on another VM to backup its drives however those drives didn't get released for some reason and when my backup job on the proxy ran it tried to backup some 2Tb volume and of course that counted as a full which then went to tape. I have corrected that by doing an exclusion of all drives except the local drives for that proxy VM. That way it will never try to backup some volume that was hot added.

So now I have another issue. That 2Tb volume that did backup using that job has created a huge backup file. Is there any way to expire off a single server of data within a vbk that includes multiple servers? I would I guess I could run an active full but that would mean doing a full backup on all of my VMs in that job. Would like to somehow shrink down my current vbk. This is doing reverse incremental right now so not sure if I can do that.

[nunciate]

Looking a little more closely. Wondering if I can use the VM Retention setting. This is under edit job, storage section, advanced button and advanced tab.
I am hoping I can set the "Remove deleted VMs data from backup after" and set that to 1 day. Then remove the proxy VM that was backed up wrong from the job. If I run the job for a couple of days I would expect it to remove the VM from my VBK completely.

Does that sound right? I am ok with removing the whole VM as doing an active full on that one will not be a problem.

[veremin]

I'd say that deleting VM and running active full backup seems to be necessary in this case.

Also, you said that source backup job runs in reversed incremental mode, right? Please be aware, that in this case each time the backup to tape job runs, it will transfer the whole .vbk file to tape medias, as it will be considered as a new one.

Thanks.

[nunciate]

We only do the synthetic full/revers incremental peace every Friday. So currently what happens is each incremental is put into a vib file. Then on Friday the roll up in the VBK and a new VRB. That seems to work very will with all of my jobs. I only get an incremental tape out during the week. We only keep those tapes for 7 days. I am running an incremental again after removing the bad VM from the job and setting the remove deleted VM after 1 day.

[veremin]

Setting deleted VM retention period to 1 won’t make your full backup smaller. So, you’d better run a new active full backup, instead. Thanks.

[nunciate]

Yea I removed the VIB files as they were a couple of tb and then I did an active full on everything in the job. Seems to be ok now.

[veremin]

Glad to hear you've nailed it down. Should any other questions arise, feel free to contact us. Thanks.