[PVE-User] PVE 3.4 - Debian 9 hang

[Eneko Lacunza]

Hi Emmanuel,

El 14/11/17 a las 10:35, Emmanuel Kasper escribió:
> On 11/14/2017 09:41 AM, Eneko Lacunza wrote:
>> Hi all,
>>
>> We have a Debian 9 VM running on a v3.4 Proxmox cluster (yes I know we
>> should upgrade it), that periodically hangs.
>>
>> When it hangs, we can't reach it neither by network nor by Console (it
>> is frezeed). Also, CPU is shown at about 52%-53% (VM has 2 cores).
>>
>> It hangs quite often, about every 1-3 days, no backups running. It is
>> running Icinga2.
>>
>> We have other Debian 9 VMs in that cluster that don't hang (but none is
>> running Icinga2)
>>
>> VM config:
>> - 512MB RAM (use tops at about 450MB)
>> - 2 cores, 1 socket (Default -kvm64)
>> - Display default
>> - ide2 - CD
>> - Hard disk scsi0 on ceph, writeback
>> - net0 e1000 (originally had virtio)
>>
>> We have tried:
>> - Changing proxmox node
>> - Changing network from virtio to e1000
>> - Increased core number from 1 to 2
>>
>> Any ideas about what can we try next?
>>
>> Thanks a lot
>> Eneko
>>
> Hi Eneko
> What is the status of the qemu process when the VM hangs ? Is it in D
> state ?
I don't think but didn't check it. Will do next time it hangs.
> Also which SCSI controller type are you using ?
It was set as default (LSI 53C895A), just changed to virtio on this 
morning's reset. I think we have changed this previously too, but can't 
say for sure.

Thanks
Eneko

-- 
Zuzendari Teknikoa / Director Técnico
Binovo IT Human Project, S.L.
Telf. 943569206
Astigarraga bidea 2, 2º izq. oficina 11; 20180 Oiartzun (Gipuzkoa)
www.binovo.es