Re: [ovirt-users] Spurious error message when the RHEV capacity is maxed out?
On Apr 24, 2014, at 07:40 , Itamar Heim <iheim(a)redhat.com> wrote:
On 04/24/2014 12:31 AM, Satya Vempati wrote:
> We have about 900 vms in a RHEV-M 3.3 installation (3.3.2-0.50.el6ev)
> with three hosts.
>
> Each VM has 1 socket, 1 core for CPU and 256 MB memory and 2 x 1TB disks
> + one 5 GB boot disk.
>
> Each host has 2 cpu sockets with 4 cpu cores per socket.
>
> When we power on about 250 of these vms, we cannot power on any more vms.
is this a 'monday morning effect' (are you trying to launch them all together)?
even that would work, we just serialize the requests so it could only take a long time...
>
> Attempting to do so gets the error (in the RHEV-M) console.
>
> Bad volume specification {'index': 0, 'iface': 'virtio',
'reqsize': '0',
> 'format': 'raw', 'bootOrder': '1',
'volumeID':
> '14aa7302-7f6d-4b80-922d-3fc090695447', 'apparentsize':
'5368709120',
> 'imageID': '97b1985e-e507-4927-91e9-7393c7777de5',
'specParams': {},
> 'readonly': 'false', 'domainID':
'0c7abccc-1b97-4b39-ab96-e977a19fac3a',
> 'optional': 'false', 'deviceId':
'97b1985e-e507-4927-91e9-7393c7777de5',
> 'truesize': '0', 'poolID':
'5849b030-626e-47cb-ad90-3ce782d831b3',
> 'device': 'disk', 'shared': 'false',
'propagateErrors': 'off', 'type':
> 'disk'}.
check both engine and vdsm logs to identify if it's on the engine side or some trouble
at the host...
Thanks,
michal
>
> This looks like a spurious error, as we can power on the same VM after
> rebooting the hosts.
>
> And also, the hosts do not really look maxed out (CPU usage is around
> 25%, memory around 50%), but no more vm’s can be powered on.
>
> I was looking for a sizing guide, and found only
> https://access.redhat.com/site/sites/default/files/attachments/rhev_sizin...
>
> Is there any other information available?
>
>
>
> _______________________________________________
> Users mailing list
> Users(a)ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>