Full Backup is not running

[pronto]

Hi Community,

I configured three backup jobs, one for each partition of our fileserver. The job for the first partition runs automatically at 08:00pm from Monday till Friday and on Saturday I configured an active full backup for each partition. In order of wasting no time, the other two partitions are configured to start the backup as soon as the first job or respective the previous job has finished. This works great during the week for the normal, incremental backups but not for the active full backup on the weekend.

The first partition starts the active full backup as configured on Saturday at 08:00pm and finished at least this weekend a few minutes after 00:00. The other partitions starts the backup as well as configured immediately but not an active full backup but a normal, incremental backup. All three jobs are configured to perform an active full backup on Saturday and it seems that Veeam did not notice that two of the three jobs had to wait until Sunday and thus simply ignored the active full backup for the following two partitions, but did a normal backup. It is noteworthy in this context that on Sunday no backup has been configured.

Will Veeam not execute pending backups if they delay in the next day? How can I prevent this incident in the future?

Thx & Bye Tom

[PTide]

Hi,

It seems that you're talking about physical server, and using Veeam Agent. Would you tell us more - is the fileserver Windows or Linux based? Also, what kind of jobs do you use - 'Managed by Agent' or 'Managed by VBR' ?

Thanks

[pronto]

Hi PTide,

It seems that you're talking about physical server, and using Veeam Agent. Would you tell us more - is the fileserver Windows or Linux based? Also, what kind of jobs do you use - 'Managed by Agent' or 'Managed by VBR' ?

Yes, the Fileserver is a physikal Server with an installed Veeam agent and it is configured for 'Managed by backup server'. I created a 'protection group' and added the fileserver in that group, it's the only server in that group. I created also an own 'backup repository' for this fileserver. The fileserver has three partitions, each contains a couple of terrabyte on data. Than I created a backup job for the first partition, with the 'protection group' I mentioned before as source and selected the first data partition (E:\) as item to backup. As target I selected the 'backup repository' I mentioned before. I enabled the option 'application aware processing' and scheduled the task from Monday till Friday at 08:00pm with an 'active full backup' on Saturday. I configured six restore points to keep.

For the other two partitions (F:\ and G:\) I created an own backup job for each partition and I used almost the same settings, only the schedule differ. To avoid wasting time, I configured the start time after the first partition has completed, respectively for the last Job after the second job has completed. With this configuration, I wanted to avoid overlapping two backup jobs and possibly running into an error because maybe the agent is blocked by the job who is already running.

The first partition, executed on Saturday, finished as expected with a full backup but it has completed on Sunday between Midnight and 01:00am. Now I run into the problem that Veeam did not execute an 'active full backup for following pending jobs but it performed an incremental job instead. If Veeam is consistent and does not do a full backup on Sunday because it was supposed to be Saturday, it should not do an incremental backup either, because there's no backup planned on Sunday. In my opinion, Veeam should have known that the jobs were delayed, because the first job did not finish in time on Saturday, but that does not seem to be the case.

If the physical agent allows multiple jobs to run concurrently, I could configure the following jobs to start with an offset of one or two hours.

Thanks for your attention & Bye Tom

[PTide]

If Veeam is consistent and does not do a full backup on Sunday because it was supposed to be Saturday, it should not do an incremental backup either, because there's no backup planned on Sunday.

I understand yorur reasoning, and I have to disagree with you. The job had been scheduled to run AFTER the first job, which it did. Setting a job to run after another job just makes it right what it says - start one after another, regardless of the day of the week. In this case, if you want your job to be inactive on Sunday, you can configure Backup Window to prevent it from running during selected period of time.

If you believe that the thing should work differently in this case, then I encourage you to leave a feature request in this very thread by describing how you'd like it to work.

Thank you

[pronto]

Hi PTide

I understand yorur reasoning, and I have to disagree with you. The job had been scheduled to run AFTER the first job, which it did. Setting a job to run after another job just makes it right what it says - start one after another, regardless of the day of the week[...]

I agree with you, only it would be great if it would then do for what it is configured, namely a full backup but it didn't. For me this is either a bug or I have to configure it somehow different. The job logically runs after the previous one - okay - but instead of performing a full backup only performing an incremental backup, does not sound logical to me.

In this case, if you want your job to be inactive on Sunday, you can configure Backup Window to prevent it from running during selected period of time.

The job should not be inactive on Sunday, rather it should do the pending full backup. Would be great, any ideas...?

Thx & Bye Tom

[PTide]

I agree with you, only it would be great if it would then do for what it is configured, namely a full backup but it didn't. For me this is either a bug or I have to configure it somehow different. The job logically runs after the previous one - okay - but instead of performing a full backup only performing an incremental backup, does not sound logical to me.

Well, Full is a property of a job, not a separate thing that just starts on its own. Every time it is time to start a job, the service basically checks it like "aha, this job has full backup configured to go off on Saturday! Is it Saturday yet? No? Then 'incremental' ". If the 1st job had finished on Saturday so the 2nd job could start on Saturday then it would be full backup trigerred.

The job should not be inactive on Sunday, rather it should do the pending full backup. Would be great, any ideas...?

Maybe there should be some sort of a 'run shifted full' checkbox, so users can choose the behaviour that fits their needs?

Meanwhile you can configure full backup to be performed on both days - Saturday and Sunday. So when the linked job starts it will always be either Saturday or Sunday and you'll get full backup no matter when the 1st job has finished.

Thanks

[pronto]

Hi PTide

Meanwhile you can configure full backup to be performed on both days - Saturday and Sunday. So when the linked job starts it will always be either Saturday or Sunday and you'll get full backup no matter when the 1st job has finished.

But what happens if the job actually gets done on Saturday? Then it would probably be executed again on Sunday? But since that will probably not be the case anyway, I can imagine that as a workaround. I can predict the duration of the first job and place it so that it runs safely until early Sunday. Or just configure only the following two jobs for Sunday (and maybe for Saturday too), then this could actually work. I think you helped me a lot, thanks a lot...

Thx & Bye Tom

[PTide]

But what happens if the job actually gets done on Saturday? Then it would probably be executed again on Sunday?

Right, it would start on Sunday. My bad, it totally slipped from my mind that you use 'Managed by VBR' type of job, therefore the logic described in my previous post does not apply here as it's VBR what starts the job, not the Agent. I apologize for misguidance.

Thanks

[pronto]

Hi PTide

Right, it would start on Sunday. My bad, it totally slipped from my mind that you use 'Managed by VBR' type of job, therefore the logic described in my previous post does not apply here as it's VBR what starts the job, not the Agent. I apologize for misguidance.

That's a pitty, it sounded really good. I have now refrained from working through the jobs one after the other. I have now set up a separate schedule for each job and configured it to be offset by two hours each. At least in case of a full backup the jobs definitely have to wait for each other now. Is there a timeout for how long they wait for each other or will they eventually cancel their execution at any time?

Thx & Bye Tom

[PTide]

Is there a timeout for how long they wait for each other or will they eventually cancel their execution at any time?

Sorry for the delay, I'm waiting to receive the exact value from the development team. The only thing that I can tell you now is that the second job will wait, and the timeout is pretty significant (24 hours or so). May I ask you why do you use three separate jobs linked one after another, why not to use one job for all disks?

Thanks