[PVE-User] vzdump backup timing out

[Emmanuel Kasper]

On 11/07/2015 10:28 PM, Michael Rasmussen wrote:
> On Sat, 7 Nov 2015 16:03:25 -0500 (EST)
> Daniel Bayerdorffer <danielb at numberall.com> wrote:
> 
>> Hi Alain, 
>>
>> Thanks in advance for your help. At first I thought my storage was full as well. So I deleted any existing backups and tried again. It worked when I did a manual backup, but again failed on a scheduled backup. 
>>
>> VM101 is about 100 Gig, and VM102 is about 9 Gig. 
>>
> I think it is caused by congestion on your backup device/network. When
> you do a manual backup there is only 1 backup running concurrently but
> when the scheduled backup runs it will start backup concurrently if the
> VM's which is to be backed up is running on different Proxmox host. To
> overcome this problem you should make a scheduled backup for each
> proxmox host running on a different time scale so that there at any
> time only runs a scheduled backup of one VM.
> 

Alternatively you can also reduce the network congestion by setting a
I/O bandwith limit in /etc/vzdump.conf.

If you backup over a Gigabit ethernet network, you can setup for instance

bwlimit: 51200

(unit is *KBytes* )

so a single backup do not try to use more than half of the network
throughput.