[Gluster-users] logfiles get flooded by warnings

[Niels de Vos]

On Fri, Sep 19, 2014 at 07:35:39PM +0530, Vijay Bellur wrote:
> On 09/19/2014 01:56 PM, Bernhard Glomm wrote:
> >Hi all,
> >
> >I'm running
> >#: glusterfs -V
> >#: glusterfs 3.4.5 built on Aug  6 2014 19:15:07
> >on ubuntu 14.04 from the semiosis ppa
> >
> >I have a replica 2 with 2 servers
> >another client does a fuse mount of a volume.
> >On rsyncing a bit of data onto the fuse mount,
> >  I get an entry like the below one on the client - for each file that
> >is copied onto the volume
> >
> >[2014-09-19 07:57:39.877806] W
> >[client-rpc-fops.c:1232:client3_3_removexattr_cbk]
> >0-<volume_name>-client-0: remote operation failed: No data available
> >[2014-09-19 07:57:39.877963] W
> >[client-rpc-fops.c:1232:client3_3_removexattr_cbk]
> >0-<volume_name>-client-1: remote operation failed: No data available
> >[2014-09-19 07:57:39.878462] W [fuse-bridge.c:1172:fuse_err_cbk]
> >0-glusterfs-fuse: 21741144: REMOVEXATTR()
> >/<path_to_file>/.<file_name_with_a_leading_dot> => -1 (No data available)
> >
> >The data itself is present and accessible on the volume and on both bricks.
> >
> >So three questions:
> >a.) what kind of data is not available? what is the client complaining
> >about?
> 
> This problem is being seen for a removexattr() operation. The client would
> have sent a request for an extended attribute to be removed from a file or
> directory. No data available/ENODATA error is seen when the file or
> directory does not have the key of the extended attribute which was
> requested to be removed. Performing strace of rsync might give a clue about
> the key of the extended attribute.
> 
> >b.) since it is a warning and the data seems to be okay - is there
> >anything I need to fix?
> 
> Usually this warning is benign in nature.
> 
> >c.) How can I get rid of the amount of log lines? it's more than 3GB/day..
> >
> 
> You can possibly have a logrotate policy to rotate based on the size of the
> log file.
> 
> I have sent across a patch to avoid flooding logs with removexattr warning
> messages when the error happens to be ENODATA or ENOATTR [1].
> 
> Regards,
> Vijay
> 
> [1] http://review.gluster.org/#/c/8781/

The patch looks ok, there just needs a bug to be filed and included in
the comments.

In the bug report (and maybe in the commit message), the log message
should get listed. This will help other users to find the bug and fix.

Could someone file a bug for this and post the bug number as a reply?
- https://bugzilla.redhat.com/enter_bug.cgi?product=GlusterFS&component=protocol

Thanks,
Niels