[ovirt-users] Volume XXX contains a apparently corrupt brick(s).

[Sahina Bose]

On 09/02/2014 02:14 AM, Fawzy Ibrhim wrote:
> Dear All;
>
> I have GlusterFS storage domain in my oVirt Engine Version: 3.4.3-1.el6.
>
> One of the peers get disconnected due to network cable unplugged and 
> after that I failed to add/remove disks from my GlusterFS domain.
>
> I have healed the GlusterFS volume and removed the corrupted index 
> file and healed the GlusterFS volume.
>
> As workaround, I have stopped all the VM and put the GlusterFS domian 
> in maintenance mode and activated it again and now I can add/remove 
> disks from the GlusterFS domain.
>
> But I still have the below warnings in my ovirt-engine.log:-
>
> 2014-09-01 23:09:47,550 WARN 
> [org.ovirt.engine.core.bll.gluster.GlusterSyncJob] 
> (DefaultQuartzScheduler_Worker-26) Volume XXX contains a apparently 
> corrupt brick(s). Hence will not add it to engine at this point.
> 2014-09-01 23:09:52,558 INFO 
> [org.ovirt.engine.core.vdsbroker.gluster.GlusterVolumesListVDSCommand] 
> (DefaultQuartzScheduler_Worker-61) START, 
> GlusterVolumesListVDSCommand(HostName = host1, HostId = 
> 0de0e7d7-a3dd-4776-9393-c36b8c4b452d), log id: 2b6381b0
> 2014-09-01 23:09:52,588 WARN 
> [org.ovirt.engine.core.vdsbroker.gluster.GlusterVolumesListReturnForXmlRpc] 
> (DefaultQuartzScheduler_Worker-61) Could not add brick 
> host2:/var/lib/exports/gluster/brick1 to volume 
> fca9d8a4-392d-4e55-a2ac-8e09a75c5cd2 - server uuid 
> d7763283-10b6-4880-a7b3-48d0353205b0 not found in cluster 
> 00000001-0001-0001-0001-00000000033a
> 2014-09-01 23:09:52,594 INFO 
> [org.ovirt.engine.core.vdsbroker.gluster.GlusterVolumesListVDSCommand] 
> (DefaultQuartzScheduler_Worker-61) FINISH, 
> GlusterVolumesListVDSCommand, return: 
> {fca9d8a4-392d-4e55-a2ac-8e09a75c5cd2=org.ovirt.engine.core.common.businessentities.gluster.GlusterVolumeEntity at 107254ac}, 
> log id: 2b6381b0
>


It looks like you're managing your gluster storage using oVirt as well. 
Since one of your hosts got disconnected (host2 ?), did the host entry 
get removed from the list of hosts in your cluster?

The gluster volume has a brick on one of the removed hosts, and it is 
unable to synchronize the brick information returned from gluster cli to 
the engine's database.



> Best Regards
> Fawzy Ibrahim
>
>
> _______________________________________________
> 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/20140902/227b008c/attachment-0001.html>