Kingsley
2016-Jun-13 15:48 UTC
[Gluster-users] 3.6.3 - how can I tell if volume is healing?
On Mon, 2016-06-13 at 18:32 +0530, Ravishankar N wrote:> On 06/13/2016 06:13 PM, Kingsley wrote: > > How can I tell whether a 3.6.34 way replicated volume is making progress > > with self healing? > > > > We did a yum update on one of the brick servers a while back, but due to > > a kernel issue the server didn't reboot, so it was down for several days > > and quite a few updates were done to the volumes before it came back up. > > > > If I look at /var/log/glusterfs/glfsheal-callrec.log there are many > > lines like this one: > > You would need to look at glustershd.log on the nodes. The > success/failures are logged there.Ah, thanks, I can see stuff ticking past slowly there. Is there any way to view or set the self-heal priority? It would be nice to speed it up a little, as it's going to take several weeks to catch up at the rate it's going. -- Cheers, Kingsley.
Ravishankar N
2016-Jun-13 16:15 UTC
[Gluster-users] 3.6.3 - how can I tell if volume is healing?
On 06/13/2016 09:18 PM, Kingsley wrote:> Is there any way to view or set the self-heal priority? It would be nice > to speed it up a little, as it's going to take several weeks to catch up > at the rate it's going.You could try changing the cluster.data-self-heal-algorithm to 'full' instead of the default 'diff' and see if that helps. Accessing the files that need heal from the client could also trigger the heals on that file. (stat </mount/filename>). -Ravi