[ovirt-users] Disabling migration when there is a problem with storage

Dan Yasny dyasny at gmail.com
Wed Feb 3 15:45:50 UTC 2016


When a storage error occurs, VMs should pause, not migrate. However, if the
VMs were io-idle, then pause (qemu EIO) might have not been triggered
before hosts started dropping into non-op.
...which triggered a migration storm for HA-enabled VMs, racing with VMs
pausing somehow (can't tell without diving into the logs for sure).

I haven't seen it happen in any of my setups, probably a new issue

On Wed, Feb 3, 2016 at 9:39 AM, Mark Steele <msteele at telvue.com> wrote:

> Hello,
>
> I'd like to know if there is a way to configure ovirt (oVirt Engine
> Version: 3.5.0.1-1.el6) to NOT migrate VM's when there is a problem with
> attached storage.
>
> Scenario - our storage array that hosts all our VM disks failed. When that
> happened, ovirt attempted to migrate running VM's from one host to another.
> Unfortunately this left many of the VM's in a 'migrating to' state that
> could only be resolved by restarting the HV's
>
> I'd prefer to have ovirt simply hang the VM's until disk IO returns.
>
> Is this possible?
>
> ***
> *Mark Steele*
> CIO / VP Technical Operations | TelVue Corporation
> TelVue - We Share Your Vision
> 800.885.8886 x128 | msteele at telvue.com | http://www.telvue.com
> twitter: http://twitter.com/telvue | facebook:
> https://www.facebook.com/telvue
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20160203/cb514fde/attachment-0001.html>


More information about the Users mailing list