[Gluster-users] gluster snapshot delete failure

Avra Sengupta asengupt at redhat.com
Fri Jul 17 09:27:45 UTC 2015


On investigating the logs attached to the bug, I see that the failure is 
in the following section:

[2015-07-16 11:26:31.663354] W [MSGID: 106073] 
[glusterd-snapshot.c:2634:glusterd_do_lvm_snapshot_remove] 0-management: 
Lvm is not mounted for brick 
Removing lv (/dev/storage_vg/e905ba76967f43efa0220c2283c87057_0).
[2015-07-16 11:26:31.684119] E [MSGID: 106044] 
[glusterd-snapshot.c:2676:glusterd_do_lvm_snapshot_remove] 0-management: 
removing snapshot of the brick 
of device /dev/storage_vg/e905ba76967f43efa0220c2283c87057_0 failed
[2015-07-16 11:26:31.684174] E [MSGID: 106044] 
[glusterd-snapshot.c:2804:glusterd_lvm_snapshot_remove] 0-management: 
Failed to remove the snapshot 
[2015-07-16 11:26:31.962264] W [MSGID: 106033] 
[glusterd-snapshot.c:2850:glusterd_lvm_snapshot_remove] 0-management: 
Failed to rmdir: /run/gluster/snaps/e905ba76967f43efa0220c2283c87057/, 
err: Directory not empty. More than one glusterd running on this node. 
[Directory not empty]

The system assumes(wrongly) that LVM is not mounted for the brick, and 
so goes ahead with deleting the brick path. But fails, because the 
directory is not empty. The same issue was happening when the original 
bug was filed, and was fixed with the fix mentioned. Would it be 
possible to have a look at your setup, so that we can get a better 
understanding of what is creating the issue, even with the fix.


On 07/17/2015 02:15 PM, Avra Sengupta wrote:
> Hi Richard,
> In reference to the issue you are facing on 
> https://bugzilla.redhat.com/show_bug.cgi?id=1232430, could you please 
> also send the output of the mount command on server 1.
> Regards,
> Avra

More information about the Gluster-users mailing list