[ovirt-users] Bad volume specification after hung migration

[Michal Skrivanek]

> On 26 Oct 2017, at 10:20, Roberto Nunin <robnunin at gmail.com> wrote:
> 
> We are running 4.0.1.1-1.el7.centos

Hi,
any reason not to upgrade to 4.1?

> 
> After a frozen migration attempt, we have two VM that after shutdown, are not anymore able to be started up again.

what do you mean by frozen? Are you talking about "VM live migration" or “live storage migration”?
How exactly did you resolve that situation, you only shut down those VMs? No other troubleshooting steps, e.g. restarting engine, hosts, things like that?

Thanks,
michal
> 
> Message returned is :
> 
> Bad volume specification {'index': '0', u'domainID': u'731d95a9-61a7-4c7a-813b-fb1c3dde47ea', 'reqsize': '0', u'format': u'cow', u'optional': u'false', u'address': {u'function': u'0x0', u'bus': u'0x00', u'domain': u'0x0000', u'type': u'pci', u'slot': u'0x05'}, u'volumeID': u'cffc70ff-ed72-46ef-a369-4be95de72260', 'apparentsize': '3221225472', u'imageID': u'3fe5a849-bcc2-42d3-93c5aca4c504515b', u'specParams': {}, u'readonly': u'false', u'iface': u'virtio', u'deviceId': u'3fe5a849bcc2-42d3-93c5-aca4c504515b', 'truesize': '3221225472', u'poolID': u'00000001-0001-0001-0001-0000000001ec', u'device': u'disk', u'shared': u'false', u'propagateErrors': u'off',u'type':u'disk'}
> 
> Probably this is caused by a wrong pointer into the database that still refer to the migration image-id.
> 
> If we search within all_disks view, we can find that parentid field isn't 00000000-0000-0000-0000-000000000000 like all other running vm, but it has a value:
> 
>        vm_names       |               parentid
> ----------------------+--------------------------------------
>  working01.company.xx | 00000000-0000-0000-0000-000000000000
>  working02.company.xx | 00000000-0000-0000-0000-000000000000
>  working03.company.xx | 00000000-0000-0000-0000-000000000000
>  working04.company.xx | 00000000-0000-0000-0000-000000000000
>  broken001.company.xx | 30533842-2c83-4d0e-95d2-48162dbe23bd <<<<<<<<<
>  working05.company.xx | 00000000-0000-0000-0000-000000000000
> 
> 
> How we can recover from this ?
> 
> Thanks in advance
> Regards,
> 
> -- 
> Robert​o​
> 
> 
> 
> _______________________________________________
> 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/20171026/e5f59ac0/attachment.html>