From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932833AbWGALia (ORCPT ); Sat, 1 Jul 2006 07:38:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932869AbWGALia (ORCPT ); Sat, 1 Jul 2006 07:38:30 -0400 Received: from tornado.reub.net ([202.89.145.182]:7336 "EHLO tornado.reub.net") by vger.kernel.org with ESMTP id S932833AbWGALi3 (ORCPT ); Sat, 1 Jul 2006 07:38:29 -0400 Message-ID: <44A65EB7.5020201@reub.net> Date: Sat, 01 Jul 2006 23:38:31 +1200 From: Reuben Farrelly User-Agent: Thunderbird 3.0a1 (Windows/20060627) MIME-Version: 1.0 To: Neil Brown CC: linux-kernel@vger.kernel.org, Andrew Morton Subject: Re: Weird RAID/SATA problem [ once was Re: 2.6.17-mm3 ] References: <20060630105401.2dc1d3f3.akpm@osdl.org> <44A5C1D5.20200@reub.net> <17573.50871.307879.557218@cse.unsw.edu.au> <44A5D079.6070505@reub.net> <17573.55937.866300.638738@cse.unsw.edu.au> <44A6390E.1030608@reub.net> <17574.15295.828980.278323@cse.unsw.edu.au> <44A64BD8.90906@reub.net> <17574.21399.979888.127483@cse.unsw.edu.au> In-Reply-To: <17574.21399.979888.127483@cse.unsw.edu.au> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On 1/07/2006 10:51 p.m., Neil Brown wrote: > On Saturday July 1, reuben-lkml@reub.net wrote: >>>> md: super_written gets error=-5, uptodate=0 >>>> >>>> messages on the console that didn't seem to want to stop... >>> '5' == EIO >>> >>> We try to write the superblock and we get EIO - something wrong somewhere. >>> >>> What sort of device are we writing to here? What controller, what >>> driver (if you know), what drives? >> The two raid-1 disks are the Seagate ST380817AS SATA disks on the onboard >> controller. The motherboard is an Intel D945GNT motherboard. See dmesg.. >> >>> Can you write to the device without using md? >> Yes. >> > > So... When md writes a superblock to this device, it reliably (or > close to reliably) gets EIO. When mkfs writes, it works fine. > > Only difference I can think of is still barriers... Does this patch > make any difference? You will be happy to know that yes, it does make a difference. Applied to -mm4, RAID-1 now comes up with all arrays in sync and everything looking good. Tried it twice, and both times raid-1 came up perfectly with md0 : active raid1 sdc2[1] sda2[0] 24410688 blocks [2/2] [UU] bitmap: 0/187 pages [0KB], 64KB chunk for each md. reuben