[PVE-User] Weird stuff happening with VM - backup failure etc

[Lindsay Mathieson]

I cloned a windows 7 VM a few days agoo, ran sysprep on it and
reactivated etc. Doubt this is related to my issues though.

Several times it has just stopped responding, Java Console abd Spice
just won't connect, no failure, they just never display anytthing.
Only thing that works is a "Stop"

Last night its backup failed:

500: Feb 12 01:38:48 INFO: creating archive
'/mnt/backup/dump/vzdump-qemu-500-2014_02_12-01_38_47.vma.lzo'
500: Feb 12 01:38:51 ERROR: unable to connect to VM 500 socket -
timeout after 31 retries
500: Feb 12 01:38:51 INFO: aborting backup job
500: Feb 12 01:48:51 ERROR: VM 500 qmp command 'backup-cancel' failed
- unable to connect to VM 500 socket - timeout after 5974 retries
500: Feb 12 01:48:52 ERROR: Backup of VM 500 failed - unable to
connect to VM 500 socket - timeout after 31 retries
500: Feb 12 01:48:52 INFO: HOOK: backup-abort snapshot 500
500: Feb 12 01:48:52 INFO: HOOK-ENV:
vmtype=qemu;dumpdir=/mnt/backup/dump;storeid=Backups;hostname=Brendan-Dev;tarfile=/mnt/backup/dump/vzdump-qemu-500-2014_02_12-01_38_47.vma.lzo;logfile=/mnt/backup/dump/vzdump-qemu-500-2014_02_12-01_38_47.log


And the VM was unresponsive this morning. I halted it, checked the
status was "Stopped" and have started a manual backup. During the
backup its state changed to "Prelaunch" and its cpu usage was at 1%-2%
Now the backup has finished, the state is back to "Stopped"


Very concerned by this erratic  behaviour.

Proxmox 3.1, non-commercial, all updates.

VM is qcow2, virtio drivers for disk and net. Storage is NFS, two node
cluster. 7 other similar VM's on this node, no problems with them.

thanks,
-- 
Lindsay