[Gluster-users] Missing files on one of the bricks

[Ravishankar N]

On 11/16/2017 04:12 PM, Nithya Balachandran wrote:
>
>
> On 15 November 2017 at 19:57, Frederic Harmignies 
> <frederic.harmignies at elementai.com 
> <mailto:frederic.harmignies at elementai.com>> wrote:
>
>     Hello, we have 2x files that are missing from one of the bricks.
>     No idea how to fix this.
>
>     Details:
>
>     # gluster volume info
>     Volume Name: data01
>     Type: Replicate
>     Volume ID: 39b4479c-31f0-4696-9435-5454e4f8d310
>     Status: Started
>     Snapshot Count: 0
>     Number of Bricks: 1 x 2 = 2
>     Transport-type: tcp
>     Bricks:
>     Brick1: 192.168.186.11:/mnt/AIDATA/data
>     Brick2: 192.168.186.12:/mnt/AIDATA/data
>     Options Reconfigured:
>     performance.cache-refresh-timeout: 30
>     client.event-threads: 16
>     server.event-threads: 32
>     performance.readdir-ahead: off
>     performance.io-thread-count: 32
>     performance.cache-size: 32GB
>     transport.address-family: inet
>     nfs.disable: on
>     features.trash: off
>     features.trash-max-filesize: 500MB
>
>     # gluster volume heal data01 info
>     Brick 192.168.186.11:/mnt/AIDATA/data
>     Status: Connected
>     Number of entries: 0
>
>     Brick 192.168.186.12:/mnt/AIDATA/data
>     <gfid:7e8513f4-d4e2-4e66-b0ba-2dbe4c803c54>
>     <gfid:9612ecd2-106d-42f2-95eb-fef495c1d8ab>
>     Status: Connected
>     Number of entries: 2
>
>     # gluster volume heal data01 info split-brain
>     Brick 192.168.186.11:/mnt/AIDATA/data
>     Status: Connected
>     Number of entries in split-brain: 0
>
>     Brick 192.168.186.12:/mnt/AIDATA/data
>     Status: Connected
>     Number of entries in split-brain: 0
>
>
>     Both files is missing from the folder on Brick1, the gfid files
>     are also missing in the .gluster folder on that same Brick1.
>     Brick2 has both the files and the gfid file in .gluster
>
>     We already tried:
>
>      #gluster heal volume full
>     Running a stat and ls -l on both files from a mounted client to
>     try and trigger a heal
>
>     Would a re-balance fix this? Any guidance would be greatly
>     appreciated!
>
>
> A rebalance would not help here as this is a replicate volume. Ravi, 
> any idea what could be going wrong here?
No, explicit lookup should have healed the file on the missing brick. 
Unless lookup did not hit afr and is served from caching translators.
Frederic, what version of gluster are you running? Can you launch 
'gluster heal volume' and see glustershd logs for possible warnings? Use 
DEBUG client-log-level if you have to.  Also, instead of stat, try a 
getfattr on the file from the mount.

-Ravi
>
> Regards,
> Nithya
>
>
>     Thank you in advance!
>
>     -- 
>     *
>     *
>     *Frederic Harmignies*
>     /High Performance Computer Administrator/
>
>     www.elementai.com <http://www.elementai.com/>
>
>     _______________________________________________
>     Gluster-users mailing list
>     Gluster-users at gluster.org <mailto:Gluster-users at gluster.org>
>     http://lists.gluster.org/mailman/listinfo/gluster-users
>     <http://lists.gluster.org/mailman/listinfo/gluster-users>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20171116/43d4f701/attachment.html>