[Gluster-users] Geo-replication has stopped processing changelogs...

[Nathan Aldridge]

Digging around into what is going on here, I may be encountering some incarnation of the rename bug (https://bugzilla.redhat.com/show_bug.cgi?id=1154307) since I have something that is doing:

* mv dir old-dir
* mkdir dir
* put lots of files back into dir (where filenames are probably the same as in old-dir)

Cheers,

Nathan

-----Original Message-----
From: Nathan Aldridge 
Sent: Monday, January 12, 2015 9:55 AM
To: 'gluster-users at gluster.org'
Subject: Geo-replication has stopped processing changelogs...

Hi,

I am running Gluster  3.6.1 on Redhat 6.5 on two boxes, using geo-replication between a master and slave volume. I have started to see the following errors:

Slave brick log:

 [2015-01-12 17:49:48.479485] W [posix-handle.c:727:posix_handle_hard] 0-bvoldata-posix: mismatching ino/dev between file /ltsa/gluster/data/bvoldata/.glusterfs/29/ee/29eea02c-3113-4078-96bc-cdea8e53848a/9/.standard.js.6ZATFL (21366602/64771) and handle /ltsa/gluster/data/bvoldata/.glusterfs/88/04/8804c0f1-ef27-4f1c-84d0-4a9c9b22d64e (21366072/64771)

Geo-rep-slave log:
[2015-01-12 17:49:48.473079] W [fuse-bridge.c:1261:fuse_err_cbk] 0-glusterfs-fuse: 120591: SETXATTR() /.gfid/b0dc2a25-5cf3-4cc0-ac6b-54b96fc21c3a => -1 (File exists)
[2015-01-12 17:49:48.478292] W [client-rpc-fops.c:240:client3_3_mknod_cbk] 0-bvoldata-client-0: remote operation failed: File exists. Path: <gfid:b0dc2a25-5cf3-4cc0-ac6b-54b96fc21c3a>/.standard.js.6ZATFL

So, geo-replication doesn't seem to be able to catch up due to these errors. There are lots of these kinds of errors. The gluster volume geo-rep ... status detail cmd yields a 'History Crawl' mode.

Any thoughts? Thanks in advance.

Cheers,

Nathan