From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750747AbWGAQ3b (ORCPT ); Sat, 1 Jul 2006 12:29:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750793AbWGAQ3b (ORCPT ); Sat, 1 Jul 2006 12:29:31 -0400 Received: from embla.aitel.hist.no ([158.38.50.22]:1163 "HELO embla.aitel.hist.no") by vger.kernel.org with SMTP id S1750747AbWGAQ3b (ORCPT ); Sat, 1 Jul 2006 12:29:31 -0400 Date: Sat, 1 Jul 2006 18:25:32 +0200 To: Andrew Morton Cc: linux-kernel@vger.kernel.org, neilb@cse.unsw.edu.au, mingo@redhat.com Subject: Re: 2.6.17-mm4 raid bugs & traces Message-ID: <20060701162532.GA14933@aitel.hist.no> References: <20060629013643.4b47e8bd.akpm@osdl.org> <20060701111153.GA10855@aitel.hist.no> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20060701111153.GA10855@aitel.hist.no> User-Agent: Mutt/1.5.11+cvs20060403 From: Helge Hafting Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org More mm4 raid-1 troubles. This time, the kernel panicked upon shutdown. I was able to write down the call trace: process swapper super_written __end_that_request_first blk_ordered_complete scsi_end_request scsi_io_completion blk_done_softirq __do_softirq call_softirq RIP:md_error RSP:ffffffff80765e00 CR2:0000000000000048 <0> kernel panic - not syncing: Aiee, killing interrupt handler Hw involved: Three raid-1, two on plain scsi and one on SATA. Each raid-1 consists of two partitions. This time, each md device was running in degraded mode. Booting into 2.6.15 in order to re-add devices and sync the RAID, I get only 2768K/sec reconstruction speed on SATA, still 1108 minutes (18 hours) to go. :-( Odd, as 2.6.17mm4 resynced this in 50min, but hit a write error (real or imagined?) immediately afterwards. The other older devices, on plain scsi, resynced much faster. 19381K/sec More than a little irritating, I need the SATA raid-1 to be in sync so lilo can install mm5 for me. 18 hours. Looks like 2.6.17mm4 doesn't like mirror devices? Helge Hafting