[Gluster-devel] Flooding of client logs with JSON fop statistics under DEBUG log-level

[Anoop C S]

Hi all,

Change #12209[1] (git hash 722ed512220395af8a707756b49df67afacda795 in
master) made fop statistics to be dumped on to client log every N
seconds. These dumps are at GF_LOG_DEBUG level with configurable dump
interval whose default value is 5 seconds. So if we happen to raise
client log-level to DEBUG, then it starts dumping the statistics.

Since we have a volume set option(diagnostics.stats-dump-interval) to
increase/decrease the dump interval can't we make its default value to
0 which will disable dumping statistics at first place?

On the other side diagnostics.stats-dump-interval
lacks OPT_FLAG_CLIENT_OPT flag[2] to notify clients on the fly about
the change in value which means that this change will take effect only
after umount and re-mount of the volume (glusterd folks can correct me
if I am wrong here).

As part of my debugging I struggled to find some other expected DEBUG
logs by scrolling these JSON dumps all over the client log. In addition
to this it eats up the disk space very quickly..!!

[1] https://review.gluster.org/#/c/12209/
[2] https://github.com/gluster/glusterfs/blob/master/xlators/mgmt/glust
erd/src/glusterd-volume-set.c#L1236

--Anoop C S.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-devel/attachments/20160319/82275b8e/attachment.html>