Backup Job not working PVE -> PBS

[cave]

Hi,

I have 2 Backup Jobs planned in my Cluster. They are basically identical, just different PBS Backend und some time Shifts.
Yesterday i checked both PBS Servers, and found the last Backup was saved at 22.03. So no backup on 5 consecutive days: 23.24.25.26.27 .... just found out by ... hmmm ... luck.

Also it happened already a few days before CET-Summer/Winter Time changes... so this one is also not related.


Yesterday i did a reboot of all Cluster Nodes, one by one. Just in Case, and waited for today 00:30 and 02:05 when the next Backup should run.
But the Backup Run for the job today did not happen.

So i Just started it now manually, to have a backup and to see if the backup-job itself works.

PBS1


PBS3

So the manual triggered run was working. both PBS Systems work fine. Backups are stored and soon get verified with the next planned verify job.

Still, i do not know why the Job is not working as expected. Both Backup Job Details look fine to me.


In the Tasklist i went back in history, to find the latest working Backup Runs and checked for failed jobs, i could not find the missing jobs. Looks like the jobs just did not start at all.

Latest run and also successful run i could find:


Matches the backups and timestamps on both PBS Systems.

Tomorrow i'll check if the backup job did run at "Next Run" or did get skipped again.

Next i'll go check date and dpkg-reconfigure tzdata on all nodes and all pbs. but i remember i set all correctly and all show same UTC and CET timestamps.
Re-Scheduling Backups to 03:00-06:00 to get away from possible date changes.


Any idea why my Backup Jobs get skipped?
Which logs to check?
Cron issue?
Which obvious piece do i miss to look at?

Related to the Upgrade to 7.4? The Timestamps match, as i upgraded on Thursday evening. (23.03.) and the next job did not start. at least it seems.

 

[fabian]

could be the pve-scheduler bug: https://bugzilla.proxmox.com/show_bug.cgi?id=4615

 

[cave]

@fabian thx, hmm, reads similar, i'll check my libev-perl version
All Nodes in the cluster are initially installed with 7.3-x. That fact does not match the bug description.

root@pve2:~# dpkg -s libev-perl
Package: libev-perl
Status: install ok installed

and

root@pve2:~# dpkg -s libpve-common-perl
Package: libpve-common-perl
Status: install ok installed

These nodes are all Debian Bullseye/11.6 to PVE 7.3 Installations (Install Proxmox VE on Debian 11 Bullseye).

 

[malakatronis]

Hello,

Same happening here. All scheduled backup jobs stopped working since the upgrade from pve-manager/7.3-6 to 7.4-3

Just upgraded to libpve-common-perl 7.3-4 from pve-no-subscription repo and restarted pvescheduler service. Now waiting for the next scheduled backup job to run to confirm it is fixed.

 

[malakatronis]

Scheduled backup jobs are working again. I can confirm the issue gets fixed by upgrading to libpve-common-perl 7.3-4.

 

[joshin]

Scheduled backup jobs are working again. I can confirm the issue gets fixed by upgrading to libpve-common-perl 7.3-4.

Thanks for this. Ran into this yesterday and today, and nice to find the fix here.