From owner-freebsd-questions  Thu Nov 19 08:55:23 1998
Return-Path: <owner-freebsd-questions@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id IAA19483
          for freebsd-questions-outgoing; Thu, 19 Nov 1998 08:55:23 -0800 (PST)
          (envelope-from owner-freebsd-questions@FreeBSD.ORG)
Received: from xxx.video-collage.com (xxx.video-collage.com [209.122.149.226] (may be forged))
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA19478
          for <questions@freebsd.org>; Thu, 19 Nov 1998 08:55:21 -0800 (PST)
          (envelope-from mi@video-collage.com)
X-Relay-IP:  
Received: (from mi@localhost)
	by xxx.video-collage.com (8.9.1a/8.9.1) id LAA13163
	for questions@freebsd.org; Thu, 19 Nov 1998 11:54:43 -0500 (EST)
Message-Id: <199811191654.LAA13163@xxx.video-collage.com>
Subject: binding fails if there is FIN_WAIT2 on this port
To: questions@FreeBSD.ORG
Date: Thu, 19 Nov 1998 11:54:43 -0500 (EST)
From: Mikhail Teterin <mi@video-collage.com>
X-Face: %UW#n0|w>ydeGt/b@1-.UFP=K^~-:0f#O:D7w<gv/&E-lL7twZCT8B~/PA4|\t$ti+22K">
	 hJ5G_<5143Bb3kOIs9XpX+"V+~$adGP:J|SLieM31VIhqXeLBli"<kcG^EOVih
	 y+z3/UR{6SCQ
X-Mailer: ELM [version 2.4ME+ PL48 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-questions@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Hello!

I'm having the following problem, but can not say for sure
if it's a bug due to my poor knowledge of TCP/IP.

Example1:

We have something like
	while (1)
	ssh -L 3129:host.on.fast.link:3128 host.on.fast.link
	sleep 3
	end

running on ttyv6 (better solutions, that will not need a tty
are welcome).

This allows to set HTTP_PROXY to localhost:3129 in, say, Netscape.

Unfortunately, when ssh dies for some reason, such as modem disconnects
for too long, the new instance of ssh can not bind to the port 3129,
because it is "in use". The only mention of it in netstat's output is
for a FIN_WAIT2 connection, which goes away when Netscape is told to
look for proxy elsewhere or to use direct connection.

Example2:

Ftp server, with ftpd being spawned from xinetd. If the ftp client (from
another host) dies, there may be a FIN_WAIT2 connection to the server's
ftp port on the client. If I try to kill and restart xinetd on the server
it will fail to bind to the ftp port. inetd does not have this problem
and is not affected by some other machine having FIN_WAIT2 connection to
its ftp port.

I'd appreciate if someone could comment on this, since I'm not sure if
this is really PR-able.

TIA!

	-mi

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message