Return-Path: Delivered-To: apmail-httpd-dev-archive@www.apache.org Received: (qmail 38360 invoked from network); 9 Mar 2004 13:16:31 -0000 Received: from daedalus.apache.org (HELO mail.apache.org) (208.185.179.12) by minotaur-2.apache.org with SMTP; 9 Mar 2004 13:16:31 -0000 Received: (qmail 36410 invoked by uid 500); 9 Mar 2004 13:16:25 -0000 Delivered-To: apmail-httpd-dev-archive@httpd.apache.org Received: (qmail 36193 invoked by uid 500); 9 Mar 2004 13:16:23 -0000 Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm Precedence: bulk Reply-To: dev@httpd.apache.org list-help: list-unsubscribe: list-post: Delivered-To: mailing list dev@httpd.apache.org Received: (qmail 36172 invoked from network); 9 Mar 2004 13:16:23 -0000 Received: from unknown (HELO castlerea.stdlib.net) (212.13.199.152) by daedalus.apache.org with SMTP; 9 Mar 2004 13:16:23 -0000 Received: from colmmacc by castlerea.stdlib.net with local (Exim 4.20) id 1B0h4D-0001gL-Lj; Tue, 09 Mar 2004 13:14:41 +0000 Date: Tue, 9 Mar 2004 13:14:41 +0000 From: Colm MacCarthaigh To: Bose.Ghanta@stratus.com Cc: "'ben@algroup.co.uk'" , "'openssl-team@openssl.org'" , dev@httpd.apache.org Subject: Re: ftp site Message-ID: <20040309131441.GA6366@castlerea.stdlib.net.> Reply-To: colm@stdlib.net References: Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: User-Agent: Mutt/1.3.28i X-Spam-Rating: daedalus.apache.org 1.6.2 0/1000/N X-Spam-Rating: minotaur-2.apache.org 1.6.2 0/1000/N On Fri, Mar 05, 2004 at 04:35:37PM -0500, Ghanta, Bose wrote: > I was working on what I originally thought was a bug in our FTP client. > Your ftp site has a very long banner (due to the crypto warnings and what > all), and the bug opened against our FTP client was that it would disconnect > partly through the login banner. After using a packet sniffer, I determined > that what is happening is that at a certain point, as your FTP server is > sending banner lines, it drops the connection. This is a relatively common failure mode for scenarios involving a stateful protocol-inspecting firewall being in the way. Many popular implementations insist on a divisional newline being within the first packet; to establish state (when using PASV) and protect against a common attack method (see below). If the banner size starts coming close to the MTU and the handshake is fragmented these implementations can break the internet. See: http://www.securityfocus.com/archive/1/46655 http://www.checkpoint.com/techsupport/alerts/pasvftp.html for a description of why the check occurs, and see: http://lists.virus.org/fw1-0302/msg00599.html for instructions on how to disable the check in the most common implementation which displays this behaviour (checkpoint). It would be worth investigating wether such a device is between you and the ftp server, and whether or not it is responsible for your problems. -- Colm MacC�rthaigh Public Key: colm+pgp@stdlib.net