[PVE-User] Backup issue

[Alain Péan]

Le 05/09/2011 15:41, Alain Péan a écrit :
> Le 05/09/2011 15:23, Muhammad Yousuf Khan a écrit :
>> i am facing an issue with a KVM container which has been incompletely 
>> backup on last Tuesday (August 30th). last week this machine was 
>> backup successfully. but this week it stuck. the process of this 
>> backup job can bee seen here.
>>
>>  Warning: bad ps syntax, perhaps a bogus '-'? See 
>> http://procps.sf.net/faq.html
>> 17514 ?        Ss     0:00      \_ /usr/bin/perl -w /usr/sbin/vzdump 
>> --quiet --node 1 --snapshot --compress --storage NFS-450 103 118 119
>> 19223 ?        S      0:00          \_ sh -c 
>> /usr/lib/qemu-server/vmtar 
>> '/mnt/pve/NFS-450/vzdump-qemu-119-2011_08_30-04_41_56.tmp/qemu-server.conf' 
>> 'qemu-server.conf' '/mnt/vzsnap0/images/119/vm-119-disk-1.raw' 
>> 'vm-disk-ide0.raw' '/mnt/vzsnap0/images/119/vm-119-disk-2.raw' 
>> 'vm-disk-ide1.raw'|cstream -t 5120000|gzip - 
>> >/mnt/pve/NFS-450/vzdump-qemu-119-2011_08_30-04_41_56.dat
>> 19224 ?        R    9358:09              \_ 
>> /usr/lib/qemu-server/vmtar 
>> /mnt/pve/NFS-450/vzdump-qemu-119-2011_08_30-04_41_56.tmp/qemu-server.conf 
>> qemu-server.conf /mnt/vzsnap0/images/119/vm-119-disk-1.raw 
>> vm-disk-ide0.raw /mnt/vzsnap0/images/119/vm-119-disk-2.raw 
>> vm-disk-ide1.raw
>> 16031 pts/0    S+     0:00          \_ grep vzdump
>>
>> this process has been stuck for a week which has been initiated on 
>> August 30th.
>>
>>
>> No error message can be seen in VZDUMP log.
>>
>> assuretyconsulting:/var/log/vzdump# cat qemu-119.log
>> Aug 30 04:41:56 INFO: Starting Backup of VM 119 (qemu)
>> Aug 30 04:41:56 INFO: running
>> Aug 30 04:41:56 INFO: status = running
>> Aug 30 04:41:57 INFO: backup mode: snapshot
>> Aug 30 04:41:57 INFO: bandwidth limit: 5000 KB/s
>> Aug 30 04:41:57 INFO: ionice priority: 7
>> Aug 30 04:41:57 INFO:   /dev/dm-2: read failed after 0 of 4096 at 0: 
>> Input/output error
>> Aug 30 04:41:58 INFO:   Logical volume "vzsnap-assuretyconsulting-0" 
>> created
>> Aug 30 04:42:00 INFO: creating archive 
>> '/mnt/pve/NFS-450/vzdump-qemu-119-2011_08_30-04_41_56.tgz'
>>
>>
>>
>> can any one tell me why the process is stuck. and what is the solution.
>>
>> Thank you,
>>
>> MYK
>>
>
> Hi Muhammad,
>
> I saw this behavior rather frequently when the backup storage was on a 
> 100 Mbit switch, with rather big VMs, > ~30GB. Since then, I upgraded 
> to a Gig switch, and it happens much less frequenly, but still 
> sometimes. I think it is a network problem.
> What is annoying is that it prevents also the backup of the remaining 
> VMs on the same host.
>
> What I do in this case is just to kill the processes involved in 
> vzdump (in your case, it would be kill 19923 and 19924).
>
> When the backup does not succeed, I see on the backup storage, beside 
> the log and tar file, a dat folder for the VM. I verify this 
> regularly, and if I see a dat file, I kill vzdump on the involved host.
>
> Alain
>

I forgot to send the answer to the list...

Just some precisions :
When I kill the vzdump processes, I see the dat folder disappear on the 
backup server, and the remaining VMs backups complete.

Alain

-- 
==========================================================
Alain Péan - LPP/CNRS
Administrateur Système/Réseau
Laboratoire de Physique des Plasmas - UMR 7648
Observatoire de Saint-Maur
4, av de Neptune, Bat. A
94100 Saint-Maur des Fossés
Tel : 01-45-11-42-39 - Fax : 01-48-89-44-33
==========================================================