From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266211AbUFPHqD (ORCPT ); Wed, 16 Jun 2004 03:46:03 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266213AbUFPHqD (ORCPT ); Wed, 16 Jun 2004 03:46:03 -0400 Received: from smtp100.mail.sc5.yahoo.com ([216.136.174.138]:24433 "HELO smtp100.mail.sc5.yahoo.com") by vger.kernel.org with SMTP id S266211AbUFPHp6 (ORCPT ); Wed, 16 Jun 2004 03:45:58 -0400 Message-ID: <40CFFAAF.7070905@yahoo.com.au> Date: Wed, 16 Jun 2004 17:45:51 +1000 From: Nick Piggin User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040401 Debian/1.6-4 X-Accept-Language: en MIME-Version: 1.0 To: Guy Van Sanden CC: Clint Byrum , linux-kernel@vger.kernel.org Subject: Re: PROBLEM: Heavy iowait on 2.6 kernels References: <1086942905.10540.69.camel@cronos.home.vsb> <1087366549.1190.6.camel@lancelot> <1087369893.11205.36.camel@cronos.home.vsb> In-Reply-To: <1087369893.11205.36.camel@cronos.home.vsb> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Guy Van Sanden wrote: > My machine is heavily used for all kinds of file serving (mainly nfs), > but also samba. > Next to that, it is my home-server, so it runs apache2 (tuned to server > only few clients), imap (cyrus), postfix, bugzilla (mysql) and distcc > (used only a few times a week). > It replaces a FreeBSD system (PII-333) running the same except distcc. > Under > > The disk system is just a regular IDE disk (udma5) (60GB) and one > external drive over USB2 (160 GB). The external drive is rather slow > (20-30 MB/sec), so I disabled it during the tests. > > The weird thing is that I see this problem too when only running bonnie. > A friend of mine tried that too under 2.6.6, his iowait went up to > 0.15%, mine to 99%. > The CPU can very easily max out the disks of course. If bonnie is doing IO to files much larger than memory, it wouldn't be surprising for io-wait to get close to 100%. Possibly your friend was doing all IO out of cache? If you definitely have a performance problem, set "Kernel Debugging" on in the "Kernel Hacking" menu, then set "Magic SysRq key" on. When your system hits this iowait problem, press Alt+SysRq+T a couple of times over a few seconds. Then post the output of `dmesg -s 1000000`. We'll see what is waiting where.