mountd fails intermittently

[Oliver Fromme]

Michael Sperber <sperber at informatik.uni-tuebingen.de> wrote:
 > Oliver Fromme <olli at lurza.secnetix.de> writes:
 > > That looks like your rpcbind(8) process died.  Can you
 > > check that with ps?  Also, are there any warnings or
 > > errors reported in /var/log/messages?
 > 
 > No, it's still running.  It shows up in rpcinfo (as does nfsd),

But mountd does not show up there?

 > and rpcbind itself also replies to pings.  In fact, I can usually
 > temporarily fix the problem by killing and restarting mountd. (Not
 > long enough to bring up my clients, though.)  (kill -HUP doesn't help,
 > btw.)  Moreover, the one client I have that's managed to mount, works.

So it's a failure in mountd.  Personally I've never seen
such a problem before, even though I'm also running NFS
servers under FreeBSD 6 (-release and -stable).

Now the one question is:  What are the circumstances under
which the problem can be reproduced?  :-)   Of course I'm
aware that that's probably a tough question.

There are a few possibilities to try to track the problem
down.

1.  First of all, it might be helpful to see the contents
    of your /etc/exports.  To be honest, I don't think that
    it is causing the problem, but you never know.

2.  Does your mountd log anything to /var/log/messages?

3.  What flags are you using with rpcbind and mountd, if
    any?  What flags are you using with the mount command
    line (i.e. anything unusual)?

4.  Please post the output from these commands (preferably
    before failure and after failure, if possible):
    # rpcinfo
    # sockstat | egrep "mountd|rpc"

5.  If all else fails, maybe tracing the mountd process
    during a failing mount attempt might be helpful.
    Personally I prefer strace (from the ports collection)
    for the more useful output, but you can also use ktrace
    which is in the base system.

Best regards
   Oliver

PS:  If you're sure it's a bug (and not PEBKAC), it might
be a good idea to submit a PR containing the information
collected so far.

-- 
Oliver Fromme,  secnetix GmbH & Co. KG, Marktplatz 29, 85567 Grafing
Dienstleistungen mit Schwerpunkt FreeBSD: http://www.secnetix.de/bsd
Any opinions expressed in this message may be personal to the author
and may not necessarily reflect the opinions of secnetix in any way.

"If you think C++ is not overly complicated, just what is a protected
abstract virtual base pure virtual private destructor, and when was the
last time you needed one?"
        -- Tom Cargil, C++ Journal