From mboxrd@z Thu Jan 1 00:00:00 1970 From: mru@kth.se (=?iso-8859-1?q?M=E5ns_Rullg=E5rd?=) Subject: Re: raid0_make_request bug: can't convert block across chunks or bigger than Date: Wed, 07 Jan 2004 11:49:31 +0100 Sender: linux-kernel-owner@vger.kernel.org Message-ID: References: <20040107023332.5ff0b9ff.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20040107023332.5ff0b9ff.akpm@osdl.org> (Andrew Morton's message: 33:32 -0800") To: Andrew Morton Cc: linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org List-Id: linux-raid.ids Andrew Morton writes: > mru@kth.se (M=E5ns Rullg=E5rd) wrote: >> >> I'm using Linux 2.6.0 on an Alpha SX164 machine. Using four ATA >> disks, hd[egik], on a Highpoint hpt374 controller, I created two ra= id0 >> arrays from hd[eg]1 and hd[ik]1, md0 and md1. From these, I create= d a >> raid1 mirror, md4, on which I created an XFS filesystem. For vario= us >> reasons, I first ran md4 with only md0 as a member and filled it wi= th >> some files, all going well. Then, I added md1, and it was synced >> properly. Now, I can mount md4 without problems. However, when I >> read things, I get this in the kernel log: >>=20 >> raid0_make_request bug: can't convert block across chunks or bigger= than 32k 439200 32 > > This was fixed post-2.6.0. 2.6.1-rc2 should be OK. I just noticed. >> raid1: Disk failure on md1, disabling device.=20 >> Operation continuing on 1 devices > > I assume this is due to the raid0 error above. I should think so. I'm still curious as to why it was only md1 that gave errors, and never md0. Was it just coincidence? --=20 M=E5ns Rullg=E5rd mru@kth.se