NFS problem (propably file lock issue)

Hi all,

I set up a NFS server, that shares the /home subtree for several clients.
After Some uptime (3 to 6 days) the clients KDE desktops refuse to start
properly: After a successful login, the screen hangs until killall X.

This behaviour lasts until the server is rebooted. Restarting NFS on the
server changes nothing.
If malfunctioning, the syslog repeats the lines, I copy/pasted below.

I read somewhere, that the problem comes with a NFS filelock issue, but I
couldn't figure out, what exactly is the problem, or where to search.

Does know a piece for my riddle?

Thanks in advance,
Clemens von Musil


Part of the /var/log/syslog:

Jul 15 09:48:48 dsasrv rpc.mountd: authenticated mount request from
dsa-102:806 for /home (/home)
Jul 15 09:49:35 dsasrv kernel: statd: server localhost not responding, timed
out
Jul 15 09:49:35 dsasrv kernel: nsm_mon_unmon: rpc failed, status=-5
Jul 15 09:49:35 dsasrv kernel: lockd: cannot monitor 192.168.123.102
Jul 15 09:51:25 dsasrv kernel: statd: server localhost not responding, timed
out
Jul 15 09:51:25 dsasrv kernel: nsm_mon_unmon: rpc failed, status=-5
Jul 15 09:51:25 dsasrv kernel: lockd: cannot monitor 192.168.123.102
Jul 15 09:53:15 dsasrv kernel: statd: server localhost not responding, timed
out
Jul 15 09:53:15 dsasrv kernel: nsm_mon_unmon: rpc failed, status=-5
Jul 15 09:53:15 dsasrv kernel: lockd: cannot monitor 192.168.123.102

Clemens von Musil wrote:

> Hi all,
>
> I set up a NFS server, that shares the /home subtree for several clients.
> After Some uptime (3 to 6 days) the clients KDE desktops refuse to start
> properly: After a successful login, the screen hangs until killall X.

Yep, locking issue. In general (I believe), it's a failed/stale lock on
~foo/.qt/qtrc that is the troublemaker.

Here's the Quick-n-dirty fix I use. for user foo having KDE lockups:
mv ~foo/.qt ~foo/.qt.BAK
cp -a ~foo/.qt.BAK ~foo/.qt
rm -rf ~foo/.qt.BAK

-- Rex