[sheepdog-users] node recovery after reboot, virtual machines slow to boot

Philip Crotwell crotwell at seis.sc.edu
Fri Oct 3 23:40:16 CEST 2014


Possible I guess, but a virsh startup on a VM is fast when the local
sheepdog is already up and not recovering. The delay is only when the
host reboots and sheepdog goes into recovery. And the issue is not
just that the VM takes time to come up, I can't even do a "virsh list"
to see if they are coming up. Certainly odd.


On Fri, Oct 3, 2014 at 11:47 AM, Andrew J. Hobbs <ajhobbs at desu.edu> wrote:
> An example from one of our running images:
>     <disk type='network' device='disk'>
>       <driver name='qemu' type='raw'/>
>       <source protocol='sheepdog' name='XXXX'/>
>       <target dev='vda' bus='virtio'/>
>     </disk>
> One thing that's going to hurt is using ide for bus type.  Virtio is much more efficient for transfer rates.  Even if you're running Windows servers in VMs, there is a RedHat signed driver for Virtio.  I wonder if perhaps that's where your delayed start is coming from?
> Here is an example libvirt device for a sheepdog disk:
>    <disk type='network' device='disk'>
>        <driver name='qemu'/>
>        <source protocol='sheepdog' name='xxxxxx'/>
>        <target dev='hda' bus='ide'/>
>      </disk>
> I do not have <host> elements, would explicitly adding multiple hosts help?
> Thanks,
> Philip
> --
> sheepdog-users mailing lists
> sheepdog-users at lists.wpkg.org
> http://lists.wpkg.org/mailman/listinfo/sheepdog-users

More information about the sheepdog-users mailing list