From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 13 Jun 2001 11:09:24 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 13 Jun 2001 11:09:14 -0400 Received: from 216-60-128-137.ati.utexas.edu ([216.60.128.137]:31872 "HELO tsunami.webofficenow.com") by vger.kernel.org with SMTP id ; Wed, 13 Jun 2001 11:09:03 -0400 Content-Type: text/plain; charset=US-ASCII From: Rob Landley Reply-To: landley@webofficenow.com To: Luigi Genoni , Ben Greear Subject: Re: Hour long timeout to ssh/telnet/ftp to down host? Date: Wed, 13 Jun 2001 06:07:49 -0400 X-Mailer: KMail [version 1.2] Cc: , In-Reply-To: In-Reply-To: MIME-Version: 1.0 Message-Id: <01061306074902.00703@localhost.localdomain> Content-Transfer-Encoding: 7BIT Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday 13 June 2001 05:40, Luigi Genoni wrote: > On Tue, 12 Jun 2001, Ben Greear wrote: > > You can tune things by setting the tcp-timeout probably..I don't > > know exactly where to set this.. > > /proc/sys/net/ipv4/tcp_fin_timeout > > default is 60. Never got that far. My problem was actually tcp_syn_retries. Remember, I was talking to a host that was unplugged. (I wasn't even getting "host unreachable" messages, the packets were just disappearing.) The default timeout in that case is rediculous do to the exponentially increasing delays between retries. 10 retries wound up being something like 20 minutes. I set it to 5 and everything works beautifully now. ssh (which retries the connection 4 times, and used to take over an hour to time out) now takes just over 3 minutes, which I can live with. Rob