From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 29 May 2002 08:56:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 29 May 2002 08:56:57 -0400 Received: from [212.176.239.134] ([212.176.239.134]:25782 "EHLO vzhik.octet.spb.ru") by vger.kernel.org with ESMTP id ; Wed, 29 May 2002 08:56:54 -0400 Message-ID: <000901c20710$486c3f40$baefb0d4@nick> Reply-To: "Nick Evgeniev" From: "Nick Evgeniev" To: "Neil Brown" Cc: In-Reply-To: <000901c206f0$cfd42620$baefb0d4@nick> <15604.45243.291277.197140@notabene.cse.unsw.edu.au> Subject: Re: 2.4.19-pre8-ac5 ide & raid0 bugs Date: Wed, 29 May 2002 16:56:42 +0400 Organization: Octet Corp. MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 X-Scanner: exiscan *17D30N-0005Es-00*7nPkbYm.xtk* http://duncanthrax.net/exiscan/ Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, Yes, I'm using lvm 1.1rc1 on top of raid0. Hardware details: dual piii box on via Apollo pro133 chipset (via694) two ATA-100 drives (FUJITSU MPG3204AH) attached to on-board controller (PROMISE 20265) & one drive attached to VIA controller. RAID0 is made on top of first two drives. 1. I don't have any problem with drive attached to via controller (in DMA33 mode) 2. The only problem that I have is with UDMA100 drives attached to promise controller... Also I have to notice that log messages about ide errors are related to drive that has no errors (according to smartctl -l, while another drive has a few records in its log) 3. I had no problems with ide with 2.4.7 kernel for a months... > > >----------------------------- > > But now I've got even more bugs in log like: > > >----------------------------- > > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block > > across chunks or bigger than 16k 37713311 4 > > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block > > across chunks or bigger than 16k 37713343 4 > > This is a request for a 4K block that is not 4K aligned... this > shouldn't happen. > Are you using LVM or something to partition the raid0 array? > ... though I seem to recall that LVM always partitions in multiples of > 4K so that shouldn't be a problem.