From: Jan Kara Subject: Re: 2.6.18-mm2: ext3 BUG? Date: Mon, 23 Oct 2006 12:41:41 +0200 Message-ID: <20061023104141.GB3162@atrey.karlin.mff.cuni.cz> References: <45257A6C.3060804@gmail.com> <20061005145042.fd62289a.akpm@osdl.org> <4525925C.6060807@gmail.com> <20061005171428.636c087c.akpm@osdl.org> <20061008063330.GA30283@lug-owl.de> <20061010070933.GE30283@lug-owl.de> <20061011104201.GD6865@atrey.karlin.mff.cuni.cz> <20061023081354.GK26271@lug-owl.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Andrew Morton , Jiri Slaby , Linux Kernel Mailing List , sct@redhat.com, adilger@clusterfs.com, linux-ext4@vger.kernel.org Return-path: Received: from atrey.karlin.mff.cuni.cz ([195.113.31.123]:13285 "EHLO atrey.karlin.mff.cuni.cz") by vger.kernel.org with ESMTP id S1751895AbWJWKky (ORCPT ); Mon, 23 Oct 2006 06:40:54 -0400 To: Jan-Benedict Glaw Content-Disposition: inline In-Reply-To: <20061023081354.GK26271@lug-owl.de> Sender: linux-ext4-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org Hello, > On Wed, 2006-10-11 12:42:02 +0200, Jan Kara wrote: > > > On Sun, 2006-10-08 08:33:30 +0200, Jan-Benedict Glaw wrote: > > > While I could reproduce it with a 200MB file, it seems I can't br= eak > > > it with a 10MB file. > > Hmm, I was running the test for several ours without any problem.= =2E. > > The kernel is 2.6.17.6, ext3 in ordered data mode, standard SATA di= sk. I'm > > now running it again and trying my luck ;). What is your testing en= vironment? >=20 > kolbe34-backup:/mnt# uname -a Thanks for info. This looks pretty similar to what I have (only that I have Athlon). > Still running Debian's 2.6.17-2-686, I'm now tracking down the file > size when I start to see this type of corruption. Right now, it seems > I never get it with a 16384 KB (16 MB) large file, but I get it with = a > 21504 KB (21 MB) file. >=20 > Is there something important that changes handling of file contents i= n > the 16..21 MB range? Umm, I've checked and found nothing obvious. We already have to use double-indirect block at 16MB, maybe reservation code does some distiction. Could you mount the filesystem with -o 'noreservation' and see whether you can still reproduce the problem? Also it may be useful to find out, whether you see the failure also with some older kernels..= =2E Honza >=20 > dumpe2fs output at http://lug-owl.de/~jbglaw/ext3-dumpe2fs.txt for > that filesystem. I'll now run with a 18.5 MB file... >=20 > MfG, JBG >=20 > --=20 > Jan-Benedict Glaw jbglaw@lug-owl.de +49-172-7= 608481 > Signature of: Wenn ich wach bin, tr=E4ume= ich. > the second : --=20 Jan Kara SuSE CR Labs