Dear All,
Following my first summary there was another good suggestion from
Richard Jackson, who told me about the lslk program available from
ftp://vic.cc.purdue.edu/pub/tools/unix/lslk/lslk.tar.gz .
This program compiles easily and enables you to list outstanding locks.
Unfortunately it didn't report any relevant ones on my machines.
Eventually I did what I had already considered but was a bit nervous
about...restarted rpc.statd and rpc.lockd on both server and client.
Restarting on the client did the trick (I was nervous because I know that
restarting NFS completely is impossible on a busy machine).
I also noticed that the client rpc.lockd was using significant amounts of
CPU time, so this may be relevant.
Regards,
Bob
Original message:
---------- Forwarded message ----------
Date: Fri, 30 Mar 2001 14:23:11 +0100 (BST)
From: Bob Vickers <bobv_at_cs.rhul.ac.uk>
Reply-To: Bob Vickers <R.Vickers_at_cs.rhul.ac.uk>
To: Tru64 Unix Managers <tru64-unix-managers_at_ornl.gov>
Subject: xmh inc processes hung
Dear All,
Overnight we had a problem with inetd, and I think there were some DNS
timeouts which had various knock-on effects. Most things started working
again when I restarted inetd, but we have a couple of users who are
died-in-the-wool xmh users and xmh hangs as soon as it tries to
incorporate new mail. It creates 'inc' processes which cannot be
killed...if you use kill -9 they become defunct.
It looks to me like a locking problem. Does anyone know a way of
clearing whatever locks inc uses? The mail spool is on an NFS file system.
Bob
==============================================================
Bob Vickers R.Vickers_at_cs.rhul.ac.uk
Dept of Computer Science, Royal Holloway, University of London
WWW:
http://www.cs.rhul.ac.uk/home/bobv
Phone: +44 1784 443691
Received on Tue Apr 03 2001 - 11:07:09 NZST