[Users] 1 node cluster problem

[René Koch (ovido)]

On Mon, 2013-04-22 at 08:53 +0200, Jakub Bittner wrote:
> Dne 22.4.2013 08:33, Itamar Heim napsal(a):
> > On 04/22/2013 09:30 AM, Jakub Bittner wrote:
> >> Hi,
> >>
> >> I have one node cluster (ovirt version 3.2) and this node went down
> >> because kernel panic. In wui I can see it is down, but VMs are still
> >> reported to be running. (they dont, ofcourse..) I manualy rebooted that
> >> host and now it is up. But problem is, that I can not activate it in wui
> >> (still red)
> >>
> >> If I press "Confirm 'Host has been Rebooted'" it says:
> >> Manual fence did not revoke the selected SPM (node1.site.com) since the
> >> master storage domain was not active or could not use another host for
> >> the fence operation.
> >>
> >> If I try to put host in maitenance it says:
> >>
> >> Error while executing action: Cannot switch the following Hosts to
> >> Maintenance mode: node1.site.com.
> >> One or more running VMs are indicated as non-migratable. The
> >> non-migratable VMs are: vm2.site.com.
> >>
> >>
> >> I removed "HA mode" from vm2.site.com, but did not do the job.
> >>
> >> I can not even Power off running hosts.
> >>
> >> Is there any way what to do?
> >>
> >> Thank you,
> >> Jakub Bittner
> >>
> >>
> >> _______________________________________________
> >> Users mailing list
> >> Users at ovirt.org
> >> http://lists.ovirt.org/mailman/listinfo/users
> >>
> >
> > iirc, there is a corner case with a single host in DC around this 
> > fixed post ovirt 3.2 (http://gerrit.ovirt.org/#/c/13045/),
> > but if you rebooted the host, and it is back up, engine shouldn't have 
> > an issue connecting to VDSM and detecting the VMs are no longer running.
> >
> > if the host/vdsm are up and running, and host is activated in engine - 
> > what status is it at?
> 
> Hypervisor node is reported to be down by ovirt engine, but it is 
> running. The VMs have question mark in their statuses.


Can you attach vdsm.log of your hypervisor?

I guess there will be error messages like missing logical volumes or
libvirt getcapability check issues. Have seen these and host-behavior
was the same as in your case. I solved it by changing storage to local
NFS. Didn't had enough time to report bugs, yet - sorry for this.


Regards,
René


> 
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users