From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267178AbUBMU2j (ORCPT ); Fri, 13 Feb 2004 15:28:39 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267182AbUBMU2j (ORCPT ); Fri, 13 Feb 2004 15:28:39 -0500 Received: from decay.franken.de ([194.94.249.92]:58050 "EHLO vivien.franken.de") by vger.kernel.org with ESMTP id S267178AbUBMU2d (ORCPT ); Fri, 13 Feb 2004 15:28:33 -0500 Date: Fri, 13 Feb 2004 17:56:59 +0100 From: Alex Goller To: Nico Schottelius Subject: Re: harddisk or kernel problem? Message-ID: <20040213165659.GA57556@vivien.franken.de> References: <20040213075403.GC1881@schottelius.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20040213075403.GC1881@schottelius.org> User-Agent: Mutt/1.4.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Feb 13, 2004 at 08:54:03AM +0100, Nico Schottelius wrote: > Freeing unused kernel memory: 140k freed > XFS mounting filesystem hda1 > Ending clean XFS mount for filesystem: hda1 > hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } > hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=8305458, > sector=8305454 These result from the ide/dma code. Those errors are normally thrown (someone correct me if i'm wrong) if ecc on that particular sector fails. A EIO error is thrown, and is handed over to the layer above (xfs in your case). These errors happen as well if you do raw reads on those sectors. > end_request: I/O error, dev hda, sector 8305454 > I/O error in filesystem ("hda3") meta-data dev hda3 block 0x776090 This is EIO handed over to xfs. > Now I am trying the following: > enabling Anticipatory I/O scheduler (additionaly to the Deadline I/O > scheduler) and disabling Vector-based interrupt indexing. I don't think this will help. The proplem is probably all about a broken sector on your disk. And if i understand the ide code correctly you can't do anything about that. > This is just a guess, can someone tell me if that is senseless or if my > harddisk is most likely broken? Yes, that's the most likely case. Smartmontools (smartctl) should show errors on the sectors mentioned in your dumps as well (they do here). Error 1689 occurred at disk power-on lifetime: 703 hours When the command that caused the error occurred, the device was active or idle After command completion occurred, registers were: ER ST SC SN CL CH DH -- -- -- -- -- -- -- 40 51 00 2c 46 58 e0 Error: UNC Commands leading to the command that caused the error were: CR FR SC SN CL CH DH DC Timestamp Command/Feature_Name -- -- -- -- -- -- -- -- --------- -------------------- 25 00 30 23 46 58 e0 00 3864.401 READ DMA EXT 25 00 32 21 46 58 e0 00 3860.425 READ DMA EXT 25 00 34 1f 46 58 e0 00 3843.845 READ DMA EXT 25 00 36 1d 46 58 e0 00 3839.939 READ DMA EXT 25 00 38 1b 46 58 e0 00 3839.914 READ DMA EXT And so on. bye, alex -- alexander goller alex@vivien.franken.de