[Gluster-users] glusterFS (v3.0.3) memory leaks?
Olivier Le Cam
Olivier.LeCam at crdp.ac-versailles.fr
Mon Mar 29 17:39:47 UTC 2010
I am about to migrate to GlusterFS but I'm experiencing some possible
memory leaks I'd like to get rid of before. This seems to be the same
problem as reported before by Larry Bates:
Setup is composed of a couple of GlusterFS v3.0.3 servers in RAID-1,
running both on Debian/Lenny.
Config files were generated with the following options:
~# glusterfs-volgen --name store1 --raid 1 gl1:/export/nunki
It's quite easy to repeat the problem just issuing an "ls -lR" on the
client: glusterfs memory usage slowly increases (it needs less than 30
minutes to acquire the whole RAM available before swap is in action):
Mem: 8193660k total, 8143884k used, 49776k free, 97156k buffers
Swap: 7815580k total, 2063112k used, 5752468k free, 726764k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
30245 root 20 0 7564m 5.3g 1116 S 1 68.4 1:06.66 glusterfs
Removing all optimization translators does not really fix the problem
although the memory usage is growing significantly slower.
I also tried echoing 3 to /proc/sys/vm/drop_caches as suggested by
Raghavendra G. ealier, with no effect at all.
Thanks in anticipation,
More information about the Gluster-users