[Gluster-users] Vol full of .*.gfs* after migrate-data

[Olivier Nicole]

Hi Amar,

>       > Is this a known bug, and is there a work-around?  If not, is
>       it safe to
>       > delete the .*.gfs* files so I can at least use the volume?
> 
> I have had the same exact behaviour with the latest package, on Ubuntu
> 11.0. The log shows a file size error after rebalance; while both file
> are identical. I presume that may be due to some caching done in the
> OS and the new file is not yet properly fulshed when it is being
> stat'ed for it's size.
> 
> 
> A patch is sent  to remove the temporary file if any error happens while
> copying the data. (ref:  http://review.gluster.com/552) 

Thank you for the patch. I have moved to test on CentOS where there is no 
such behaviour.

> Mean time, if you send me the exact log of 'rebalance process' (found at
> /var/log/glusterfs/etc-glusterd-mount-<volname>.log) it will help to corner
> the issue faster. Anyways, We will look into possibilities of the
> mismatching file size errors.

I attached a bzip'ed/tar'ed file that contains the log of 
etc-glusterd-mount-test-vol.log and nfs.log (that shows the error).

Best regards,

Olivier
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gluster.tar.bz2
Type: application/octet-stream
Size: 170691 bytes
Desc: 
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20111003/04d3768a/attachment.obj>