From owner-freebsd-isp Tue Jan 28 15:50:40 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id PAA19833 for isp-outgoing; Tue, 28 Jan 1997 15:50:40 -0800 (PST) Received: from root.com (implode.root.com [198.145.90.17]) by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id PAA19806; Tue, 28 Jan 1997 15:50:34 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by root.com (8.7.6/8.6.5) with SMTP id PAA17739; Tue, 28 Jan 1997 15:48:18 -0800 (PST) Message-Id: <199701282348.PAA17739@root.com> X-Authentication-Warning: implode.root.com: Host localhost [127.0.0.1] didn't use HELO protocol To: Michael Smith cc: shovey@buffnet.net (Steve), terry@lambert.org, robert@nanguo.chalmers.com.au, freebsd-questions@freebsd.org, freebsd-isp@freebsd.org, hackers@freebsd.org Subject: Re: progress report on connection problems In-reply-to: Your message of "Wed, 29 Jan 1997 09:13:31 +1030." <199701282243.JAA13081@genesis.atrad.adelaide.edu.au> From: David Greenman Reply-To: dg@root.com Date: Tue, 28 Jan 1997 15:48:18 -0800 Sender: owner-isp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >Steve stands accused of saying: >> > >> > 1) Contact the vendor for a fix; one probably exists. >> > >> > 2) If no fix is available, turn extension off on the FreeBSD >> > system, and submit a bug report to the vendor so that a >> > fix will happen. >> >> Turning extensions off does not stop the problem. > >Ah, but it does. It's just that other FreeBSD systems (like Yahoo) still >have them turned on, so the Annex is dropping its guts regardless. > >As has been mentioned, the _correct_ response is to tell your provider >that their Annex is _BROKEN_, and that they need to upgrade its >software (which I believe is a free exercise, but you can ransack >Xylogics' website for details there) preferably before you take your >business elsewhere. The traces that have been posted don't show that as being the problem. TCP extensions cause problems at connection startup and the traces are showing that a regular data packet is being dropped. One possible reason that this problem is being noticed when FreeBSD machines are involved might be due to FreeBSD's Path MTU Discovery causing the packets to be large. It seems likely to me that the Annex can't deal with large packets some or all of the time and drops the packet rather than fragmenting and/or without sending the proper ICMP message (which would break MTU discovery). -DG David Greenman Core-team/Principal Architect, The FreeBSD Project