Backup no longer works after upgrade to 2.3

[Nico Haase]

Hi there,
I upgraded to Proxmox 2.3 yesterday, and today I received a status mail claiming that the backup didn't work. This is the detailed log:

vzdump 104 --quiet 1 --mailto ninja@D120.de --mode snapshot --compress lzo --storage ninjaMounted --node proxmox

104: Mar 05 14:00:02 INFO: Starting Backup of VM 104 (qemu)
104: Mar 05 14:00:02 INFO: status = running
104: Mar 05 14:00:02 INFO: backup mode: snapshot
104: Mar 05 14:00:02 INFO: ionice priority: 7
104: Mar 05 14:00:02 INFO: creating archive '/mnt/pve/ninja/dump/vzdump-qemu-104-2013_03_05-14_00_02.vma.lzo'
104: Mar 05 14:00:02 ERROR: Parameter 'backupfile' is missing
104: Mar 05 14:00:02 INFO: aborting backup job
104: Mar 05 14:00:02 ERROR: VM 104 qmp command 'backup-cancel' failed - The command backup-cancel has not been found
104: Mar 05 14:00:03 ERROR: Backup of VM 104 failed - Parameter 'backupfile' is missing

The backup was running using LVM snapshots until yesterday, but as this has changed (or haven't it...?) the backup jobs should have been updated through the upgraded...?

 

[dietmar]

The new backup uses new qemu features, so you need to stop, then re-start the VM to make it work.

 

[Nico Haase]

Is this documented anywhere? Why can't the backup run the old way unto the next reboot? I can be happy to have only two VMs and not more, and these are also uncritical to reboot....

 

[tom]

Is this documented anywhere? Why can't the backup run the old way unto the next reboot? I can be happy to have only two VMs and not more, and these are also uncritical to reboot....

yes, the upgrade howto is always the same and includes the hint that you should stop your VM´s.

See http://pve.proxmox.com/wiki/Downloads#Update_a_running_Proxmox_Virtual_Environment_2.x_to_latest_2.3

 

[Cayuga]

I discovered that you can get around this problem with out stopping/restarting the VM. Just migrate it to another node (and migrate back if you want).

 

[ozgurerdogan]

I discovered that you can get around this problem with out stopping/restarting the VM. Just migrate it to another node (and migrate back if you want).

Only stop and srart vm is enough? Why you mention about migration?

 

[Cayuga]

In my usage here, I've been able to get around the problem about 40 times by migrating machines rather than stopping/starting.

 

[ozgurerdogan]

Why do not you try stop / start as an easier way?

 

[Cayuga]

Because downtime wasn't a good option for us and we wanted the vm's backed up.