[ovirt-users] post glusterfs 3.4 -> 3.5 upgrade issue in ovirt (3.4.0-1.fc19): bricks unavailable

[Alastair Neil]

I just did a rolling upgrade of my gluster storage cluster to the latest
3.5 bits.  This all seems to have gone smoothly and all the volumes are on
line.  All volumes are replicated 1x2

The ovirt console now insists that two of my volumes , including the
vm-store  volume with my vm's happily running have no bricks up.

It reports "Up but all bricks are down"

This would seem to be impossible.  Gluster  on the nodes itself reports no
issues

[root at gluster1 ~]# gluster volume status vm-store
> Status of volume: vm-store
> Gluster process Port Online Pid
>
> ------------------------------------------------------------------------------
> Brick gluster0:/export/brick0/vm-store 49158 Y 2675
> Brick gluster1:/export/brick4/vm-store 49158 Y 2309
> NFS Server on localhost 2049 Y 27012
> Self-heal Daemon on localhost N/A Y 27019
> NFS Server on gluster0 2049 Y 12875
> Self-heal Daemon on gluster0 N/A Y 12882
>
> Task Status of Volume vm-store
>
> ------------------------------------------------------------------------------
> There are no active volume tasks



As I mentioned the vms are running happily
initially the ISOs volume had the same issue.  I did a volume start and
stop on the volume as it was not being activly used and that cleared up the
issue in the console.  However, as I have VMs running  I can't so this for
the vm-store volume.


Any suggestions, Alastair
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20140521/fb1c10c0/attachment-0001.html>