[Gluster-users] Corrupted geo-replication CHANGELOG after running out of space in LVM thin pool

Tanner Bruce tanner.bruce at farmersedge.ca
Wed Aug 23 16:33:26 UTC 2017


I have a gluster cluster running with geo-replication. The volume that is being geo-replicated is running on a LVM thin pool. The thin pool overflowed causing a crash. I extended the thin pool LV and remounted which brought the volume back online and healthy. When I try to restart the geo-replication, I'm having the bricks that overflowed go to the Faulty state. The changes.log shows [gf_changelog_consume_wrap] 0-gfchangelog: could not parse changelog ...brick/.glusterfs/changelogs/CHANGELOG.1502410242 along with a few other changelogs.

I would like to know:

1) can I restore/fix those changelogs, or

2) can I simply nuke those changelogs (with the possibility of some data not being replicated [that's somewhat OK]



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20170823/dd3e821e/attachment.html>

More information about the Gluster-users mailing list