From owner-freebsd-stable  Fri Nov  8 04:47:32 1996
Return-Path: owner-stable
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id EAA00908
          for stable-outgoing; Fri, 8 Nov 1996 04:47:32 -0800 (PST)
Received: from time.cdrom.com (time.cdrom.com [204.216.27.226])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id EAA00902;
          Fri, 8 Nov 1996 04:47:26 -0800 (PST)
Received: (from jkh@localhost) by time.cdrom.com (8.8.2/8.6.9) id EAA01694; Fri, 8 Nov 1996 04:43:00 -0800 (PST)
Date: Fri, 8 Nov 1996 04:43:00 -0800 (PST)
From: "Jordan K. Hubbard" <jkh@time.cdrom.com>
Message-Id: <199611081243.EAA01694@time.cdrom.com>
To: wpaul@freebsd.org
Subject: Weirdie in mountd perhaps you can help me with...
Cc: dfr@freebsd.org, stable@freebsd.org
Sender: owner-stable@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

As has been reported by several people, including myself, there's still
some problem with "NFS" in 2.1.5 where essentially mountd "goes away" and
refuses to answer further requests.  Sometimes it goes away entirely, in fact,
and has to be restarted.

In order to try and get to the bottom of this one, I compiled up a version
of mountd with debugging and attached to it with gdb right after the system
came up.  I've been running this way all day, and just now reproduced one
of the failures, which was the "long hang" (I haven't had it outright die yet,
thought it's been mysteriously HUP'd twice by the system when an *outgoing*
mount request was made, and for no reason that I can see).  The debugger
shows the following stack trace:

#0  0x27345 in select ()
#1  0x18bbc in res_send ()
#2  0x168c7 in res_query ()
#3  0x1057a in _gethostbydnsaddr ()
#4  0xf6c0 in gethostbyaddr ()
#5  0x1d2d in mntsrv (rqstp=0xefbfda1c, transp=0x4d080)
    at /usr/src/sbin/mountd/mountd.c:413
#6  0xa082 in svc_getreqset ()
#7  0x48db in svc_run ()
#8  0x1a39 in main (argc=1, argv=0xefbfdadc)
    at /usr/src/sbin/mountd/mountd.c:317

Where it appears to be hanging in res_send()'s select.  This, after a long
interval, finally does time out and drops us back to mountd's usual select,
e.g.:

#0  0x27345 in select ()
#1  0x48c9 in svc_run ()
#2  0x1a39 in main (argc=1, argv=0xefbfdadc)
    at /usr/src/sbin/mountd/mountd.c:317

The corresponding mount request also succeeds at the end of this timeout.

Any ideas?

						Jordan