Gluster users - Jun 2014 - Glusterfs 3.4.4 - memory leak in fuse_getxattr

Hello,

while evaluating Gluster, I quickly noticed abnormal memory usage of 
glusterfs 3.4.4 FUSE client. According to valgrind, there's a memory 
leak in fuse_getxattr:

==31375== 1,145,000 (1,143,168 direct, 1,832 indirect) bytes in 624 
blocks are definitely lost in loss record 348 of 354
==31375==    at 0x4C2A554: calloc (in 
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==31375==    by 0x4E70C4D: __gf_calloc (in 
/usr/lib64/libglusterfs.so.0.0.0)
==31375==    by 0x797E396: get_fuse_state (in 
/usr/lib64/glusterfs/3.4.4/xlator/mount/fuse.so)
==31375==    by 0x7983333: fuse_getxattr (in 
/usr/lib64/glusterfs/3.4.4/xlator/mount/fuse.so)
==31375==    by 0x7998737: fuse_thread_proc (in 
/usr/lib64/glusterfs/3.4.4/xlator/mount/fuse.so)
==31375==    by 0x54EFF39: start_thread (pthread_create.c:308)

This memory leak seems to be addressed by 
http://review.gluster.org/#/c/5392/, but the patch is not included in 
stable 3.4.x series. I applied it manually to 3.4.4 and confirm that 
it fixes the problem.

Note that the patch is 11 months old. Why it isn't merged in 3.4 
branch yet?


Best regards,

Martin Svec

On 24/06/2014, at 5:35 PM, Martin Svec wrote:
<snip>
> Note that the patch is 11 months old. Why it isn't merged in 3.4 branch
yet?

Sounds like we just need to backport it to release-3.4 branch then.

+ Justin

--
GlusterFS - http://www.gluster.org

An open source, distributed file system scaling to several
petabytes, and handling thousands of clients.

My personal twitter: twitter.com/realjustinclift