From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sat, 19 May 2001 03:04:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sat, 19 May 2001 03:04:38 -0400 Received: from skif.spylog.com ([194.67.35.250]:40515 "HELO skif.spylog.com") by vger.kernel.org with SMTP id ; Sat, 19 May 2001 03:04:28 -0400 Date: Sat, 19 May 2001 11:04:12 +0400 From: Peter Zaitsev X-Mailer: The Bat! (v1.51) Reply-To: Peter Zaitsev Organization: http://www.spylog.ru X-Priority: 3 (Normal) Message-ID: <195243776381.20010519110412@spylog.ru> To: linux-kernel@vger.kernel.org Subject: Linux RAID5 issues. MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello linux-kernel, I'm using software raid5 on about 30 servers, and Yet twice I had a serious data loss becouse of the behavior of linux RAID device. In several cases I've got more then one of drives completely disconnected. I have no ideas why this happened but this had something to do with new AHA controler drivers and enabled TCQ. After reboot all drivers turned back. Other options for this to happen may be a controler failure, then array is on several of them. The linux RAID behavior in this case was to still allow write to the array, so what only writes to online drives there made. This have produced a huge mess on the drive after half an hour of running in this mode. The better solution I think would be block writes to the array after second drive fail. This would at least give users more recovery options. -- Best regards, Peter mailto:pz@spylog.ru