From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933017AbWFZUN5 (ORCPT ); Mon, 26 Jun 2006 16:13:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S933013AbWFZUN4 (ORCPT ); Mon, 26 Jun 2006 16:13:56 -0400 Received: from 1wt.eu ([62.212.114.60]:20489 "EHLO 1wt.eu") by vger.kernel.org with ESMTP id S933018AbWFZUNz (ORCPT ); Mon, 26 Jun 2006 16:13:55 -0400 Date: Mon, 26 Jun 2006 22:01:20 +0200 From: Willy Tarreau To: morrowc+kernel@ops-netman.net Cc: linux-kernel@vger.kernel.org Subject: Re: PROBLEM: crashes of 2.4.31 kernel (oops included) Message-ID: <20060626200120.GK13255@w.ods.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.10i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello, On Mon, Jun 26, 2006 at 05:52:37AM +0000, morrowc+kernel@ops-netman.net wrote: > >From URL - http://kernel.org/pub/linux/docs/lkml/reporting-bugs.html > > [1.] One line summary of the problem: kernel panic/oops system hangs > [2.] Full description of the problem/report: > After some random period of time (from 1 hour to several days) kernel > receives an 'oops' and dies. Requires power cycle to recover. I've a few > of the 'oops' output saved and converted with ksymoops if required > (included one below) Error appear to be in the iptables > connection_tracking actually. > > [3.] Keywords (i.e., modules, networking, kernel): > networking kernel oops > [4.] Kernel version (from /proc/version): > Linux version 2.4.31 (root@neo-u2.ops-netman.net) (gcc version 3.3.5 > (Debian 1:3.3.5-13)) #1 SMP Sun Mar 12 15:09:22 GMT 2006 Many bugs have been fixed in netfilter code since 2.4.31, you should really upgrade, either to 2.4.33-rc2, or to 2.4.31-hf32.6 which contains adds to 2.4.31 all relevant fixes till two weeks ago. You can find it here : http://linux.exosec.net/kernel/2.4-hf/ If the problem is new, it can indicate a hardware problem, or someone succeeding in exploiting one of those old bugs. Regards, Willy