[PVE-User] vzdump backups are randomly failing

[Laurent CARON]

On Wed, Oct 24, 2012 at 08:58:15AM +0000, Dietmar Maurer wrote:
> > I'm having trouble (randomly) while backing up VMs with the following error:
> > 
> > 616: Oct 24 01:29:49 INFO: Starting Backup of VM 616 (qemu)
> > 616: Oct 24 01:29:49 INFO: status = running
> > 616: Oct 24 01:29:50 INFO: backup mode: suspend
> > 616: Oct 24 01:29:50 INFO: ionice priority: 7
> > 616: Oct 24 01:29:51 INFO: suspend vm
> > 616: Oct 24 01:29:52 INFO: trying to aquire lock... failed
> 
> Another vzdump is already running.

Hi Dietmar,

It then means something strange is happening (please see below):

615: Oct 24 01:22:32 INFO: Starting Backup of VM 615 (qemu)
615: Oct 24 01:22:32 INFO: status = stopped
615: Oct 24 01:22:33 INFO: backup mode: stop
615: Oct 24 01:22:33 INFO: ionice priority: 7
615: Oct 24 01:22:34 INFO: creating archive '/mnt/pve/backups_delphi/dump/vzdump-qemu-615-2012_10_24-01_22_32.tar.lzo'
615: Oct 24 01:22:34 INFO: adding '/mnt/pve/backups_delphi/dump/vzdump-qemu-615-2012_10_24-01_22_32.tmp/qemu-server.conf' to archive ('qemu-server.conf')
615: Oct 24 01:22:34 INFO: adding '/dev/VG_FAS-2040-SIEGE-NAGIOS-SIEGE-001/vm-615-disk-1' to archive ('vm-disk-ide0.raw')
615: Oct 24 01:29:48 INFO: Total bytes written: 21474839040 (47.19 MiB/s)
615: Oct 24 01:29:48 INFO: archive file size: 1.68GB
615: Oct 24 01:29:48 INFO: delete old backup '/mnt/pve/backups_delphi/dump/vzdump-qemu-615-2012_10_20-01_38_13.tar.lzo'
615: Oct 24 01:29:49 INFO: Finished Backup of VM 615 (00:07:17)

616: Oct 24 01:29:49 INFO: Starting Backup of VM 616 (qemu)
616: Oct 24 01:29:49 INFO: status = running
616: Oct 24 01:29:50 INFO: backup mode: suspend
616: Oct 24 01:29:50 INFO: ionice priority: 7
616: Oct 24 01:29:51 INFO: suspend vm
616: Oct 24 01:29:52 INFO: trying to aquire lock... failed
616: Oct 24 01:29:52 INFO: can't lock file '/var/log/pve/tasks/.active.lock' - can't aquire lock - Interrupted system call
616: Oct 24 01:29:52 ERROR: Backup of VM 616 failed - command 'qm suspend 616 --skiplock' failed: exit code 4

The backups are completed normally until this error appears. Since this
is the backup job run via cron, maybe the next backup is starting too
early before the previous one is finished.

Is there a way I can delay the next backup by a few seconds after the
previous one finishes ?

Extract from :/etc/cron.d/vzdump:
30 21 * * 2,5,7       root vzdump --quiet 1 --mode suspend --mailto admin-proxmox at lncsa.fr --all 1 --node proxmox-siege-001 --compress 1 --maxfiles 2 --storage backups_delphi

Thanks