From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964867AbWLMMJT (ORCPT ); Wed, 13 Dec 2006 07:09:19 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964885AbWLMMJT (ORCPT ); Wed, 13 Dec 2006 07:09:19 -0500 Received: from ogre.sisk.pl ([217.79.144.158]:57451 "EHLO ogre.sisk.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964867AbWLMMJS (ORCPT ); Wed, 13 Dec 2006 07:09:18 -0500 X-Greylist: delayed 1200 seconds by postgrey-1.27 at vger.kernel.org; Wed, 13 Dec 2006 07:09:18 EST From: "Rafael J. Wysocki" To: Neil Brown Subject: Re: 2.6.19-mm1 (md/raid1 randomly drops partitions - possible sata_uli problem) Date: Wed, 13 Dec 2006 12:51:19 +0100 User-Agent: KMail/1.9.1 Cc: Andrew Morton , linux-kernel@vger.kernel.org References: <20061211005807.f220b81c.akpm@osdl.org> <200612122155.42213.rjw@sisk.pl> <17791.20220.200213.305946@cse.unsw.edu.au> In-Reply-To: <17791.20220.200213.305946@cse.unsw.edu.au> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200612131251.20256.rjw@sisk.pl> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday, 13 December 2006 01:53, Neil Brown wrote: > On Tuesday December 12, rjw@sisk.pl wrote: > > > > > > So when md writes to write out the superblock, to gets EIO... Odd that > > > you aren't getting errors for normal writes. > > > > > > What devices are the md/raid1 built on? > > > > Sata drives, on sata_uli. > > > > > > > > > > I'll try to reproduce it tomorrow and collect some more information. > > > > > > Thanks. More information is definitely better than less, so send over > > > anything you can find. > > > > Okay, seems to be readily reproducible, dmesg output from the failing kernel > > attached. > > Weird. You are getting silent write errors... > > Can you write to these drives are all? e.g. > > dd if=/dev/sdb3 of=/tmp/tmp count=1 > dd if=/tmp/tmp of=/dev/sdb3 oflag=direct > > (hopefully 'direct' will cause write errors to be passed up). Unfortunately I have no access to the machine right now. > I really think this looks like a sata problem, not an md problem. That's possible, but everything except for the md RAID seems to work. Strange. I think I'll wait until the next -mm is out and check if the problem goes away. ;-) Greetings, Rafael -- If you don't have the time to read, you don't have the time or the tools to write. - Stephen King