oom-killer kills almost all due to nfs-kernel-server bug?

[Vladimir]

From time to time (sometimes it happens once a week, sometimes - several times a day) on loaded fileserver oom-killer began to kill almost everything. The only way to stop this is to restart the nfs-kernel-server (if I'm lucky enough to be able to log in using ssh).
A sample of such "kills" is attached as file named syslog (the last line is a manual reboot of this server - it was impossible to restart nfs-kernel-server this time).
From the other hand, it soes not seems that server is out of memory: it has free RAM and a lot of swap space:

15:15:02 kbmemfree kbmemused %memused kbbuffers kbcached kbcommit %commit
15:15:02 471904 3588184 88,38 15672 2244164 577396 5,82
15:25:01 472852 3587236 88,35 4356 2239092 576804 5,81
15:35:02 711904 3348184 82,47 3896 2271532 204888 2,06
Average 681650 3378438 83,21 162473 2151526 502496 5,06

15:42:38 LINUX RESTART

(this is last incident)

I have sar running, but I could not understand the reason of such oom-killer's behavior.

If there is any need of additional information - I would be happy to send it, but please note that this is a production server, so I could not reboot it too many times.

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: nfs-kernel-server 1:1.2.0-4ubuntu4.1
ProcVersionSignature: Ubuntu 2.6.32-33.72-generic-pae 2.6.32.41+drm33.18
Uname: Linux 2.6.32-33-generic-pae i686
Architecture: i386
Date: Wed Sep 21 16:15:01 2011
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release i386 (20100427.1)
ProcEnviron:
 LANG=ru_RU.UTF-8
 SHELL=/bin/bash
SourcePackage: nfs-utils

[Vladimir (vladimir-kozlov)]

Found the following thread: http://linux-nfs.org/pipermail/nfsv4/2010-January/012042.html
It seems that this is the reason - I have very similar situation here...

[dino99 (9d9)]

Closing that outdated report as EOL has been reached long time ago