From owner-freebsd-bugs  Sat Dec  6 11:03:50 1997
Return-Path: <owner-freebsd-bugs>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id LAA06395
          for bugs-outgoing; Sat, 6 Dec 1997 11:03:50 -0800 (PST)
          (envelope-from owner-freebsd-bugs)
Received: from kithrup.com (kithrup.com [205.179.156.40])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id LAA06390
          for <bugs@freebsd.org>; Sat, 6 Dec 1997 11:03:47 -0800 (PST)
          (envelope-from sef@kithrup.com)
Received: (from sef@localhost)
	by kithrup.com (8.8.8/8.8.7) id LAA24325;
	Sat, 6 Dec 1997 11:03:46 -0800 (PST)
	(envelope-from sef)
Date: Sat, 6 Dec 1997 11:03:46 -0800 (PST)
From: Sean Eric Fagan <sef@kithrup.com>
Message-Id: <199712061903.LAA24325@kithrup.com>
To: spork@super-g.com
Subject: Re: Is this a bug?
Cc: bugs@freebsd.org
Sender: owner-freebsd-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>I've seen this many times; if you don't properly umount the nfs stuff and
>a machine disappears, it's hell to get them talking again.  

The thing is, I don't think it's just NFS related.  I think it's TCP related
-- specifically, the problem seemed to be when the rebooted machine tried to
establish a connection to the non-rebooted machine, where the new <src addr,
src port>:<dst addr, dst port> matched the old ones.  This is not quite
right behaviour -- for the rlogin I tried, the non-rebooted machine should
have realized that the connection was gone.  Or I think it should have,
anyway :).

This affected me with NFS, because I am using TCP for NFS -- and there was
no way (that I could see) to cause the connection to be closed on the
non-rebooted machine.  (For the rlogin, I did a 'kill -HUP' of the shell on
the defunct connection; that caused the connection to finally close, which
meant I could log in again.)

Sean.