[Gluster-users] Memory leak in gluster 5.4

[Sanju Rakonde]

Hi Christian,

I see below errors when I try to unzip the file.

[root at localhost Downloads]# unzip gluster_coredump.zip
Archive:  gluster_coredump.zip
checkdir error:  coredump exists but is not directory
                 unable to process coredump/.
checkdir error:  coredump exists but is not directory
                 unable to process
coredump/core.glusterd.0.ed02597e2d374210985795ab82dd48e7.2209.1557381154000000.lz4.
checkdir error:  coredump exists but is not directory
                 unable to process
coredump/core.glusterfsd.0.ed02597e2d374210985795ab82dd48e7.2634.1557381672000000.lz4.
checkdir error:  coredump exists but is not directory
                 unable to process
coredump/core.glusterfsd.0.ed02597e2d374210985795ab82dd48e7.2653.1557381626000000.lz4.
[root at localhost Downloads]#

Periodic statedumps will be much helpful in debugging memory leaks than
coredumps.

Thanks,
Sanju

On Thu, May 16, 2019 at 2:57 PM Christian Meyer <chrmeyer at chrmeyer.de>
wrote:

> Hi everyone!
>
> I'm using a Gluster 5.4 Setup with three Nodes and three volumes
>  (one is the gluster shared storage). The other are replicated volumes.
> Each node has 64GB of RAM.
> Over the time of ~2 month the memory consumption of glusterd grow
> linear. An the end glusterd used ~45% of RAM the brick processes
> together ~43% of RAM.
> I think this is a memory leak.
>
> I made a coredump of the processes (glusterd, bricks) (zipped ~500MB),
> hope this will help to find the problem.
>
> Could someone please have a look on it?
>
> Download Coredumps:
> https://s3.eu-central-1.amazonaws.com/glusterlogs/gluster_coredump.zip
>
> Kind regards
>
> Christian
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> https://lists.gluster.org/mailman/listinfo/gluster-users
>


-- 
Thanks,
Sanju
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20190529/87009ea6/attachment.html>