On 05 May 2016, at 16:18, Bhaskarakiran <byarlaga@redhat.com> wrote:

Hi,

I have a HCI setup running on 3 nodes and created 6 VM's. Was running IO (like dd and linux untar) on those VM's overnight. Next day i saw that for 2 of the nodes vdsmd and libvirtd services failed and if manually started, they don't come up. All the VM's state has changed to 'unknown' and failed to migrate. Can someone help looking at the logs and figure out the RCA. Let me know what all logs are needed, i can post the same.

Hi,

I saw your other bugs you reported already. In all cases it doesn’t seem like ovirt’s fault (except the buggy vdsm recovery flow), the underlying reason is that something got broken in either libvirt or qemu.

For that you better enable libvirt debug logging as the default level is not so useful.

You can find more details about logging at http://www.ovirt.org/develop/developer-guide/vdsm/log-files/

Once you have that please share/send vdsm.log (it’s rotated often, so check the times to cover the time from VM creation all the way to failure), libvirt.log with debug info and VM’s qemu log from /var/log/libvirt/qemu/<vm name>.log

Thanks,

michal

Thanks,

Bhaskarakiran.

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