From owner-freebsd-questions@FreeBSD.ORG Wed Sep 28 11:08:18 2005 Return-Path: X-Original-To: freebsd-questions@freebsd.org Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1457116A41F for ; Wed, 28 Sep 2005 11:08:18 +0000 (GMT) (envelope-from sandy@krvarr.bc.ca) Received: from szamoca.krvarr.bc.ca (szamoca.krvarr.bc.ca [142.179.111.232]) by mx1.FreeBSD.org (Postfix) with ESMTP id 88FEC43D48 for ; Wed, 28 Sep 2005 11:08:15 +0000 (GMT) (envelope-from sandy@krvarr.bc.ca) Received: from szamoca.krvarr.bc.ca (localhost [127.0.0.1]) by szamoca.krvarr.bc.ca (8.13.1/8.12.11) with ESMTP id j8SB8DCw032576; Wed, 28 Sep 2005 04:08:13 -0700 (PDT) (envelope-from sandy@szamoca.krvarr.bc.ca) Received: (from sandy@localhost) by szamoca.krvarr.bc.ca (8.13.1/8.12.11/Submit) id j8SB8DgH032573; Wed, 28 Sep 2005 04:08:13 -0700 (PDT) (envelope-from sandy) From: Sandy Rutherford MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <17210.31132.778408.755267@szamoca.krvarr.bc.ca> Date: Wed, 28 Sep 2005 04:08:12 -0700 To: Michael Conlen In-Reply-To: References: X-Mailer: VM 7.19 under Emacs 21.3.1 X-krvarr.bc.ca-MailScanner-Information: Please contact postmaster@krvarr.bc.ca for more information. X-krvarr.bc.ca-MailScanner: Not scanned: please contact postmaster@krvarr.bc.ca for details. X-krvarr.bc.ca-MailScanner-From: sandy@szamoca.krvarr.bc.ca Cc: FreeBSD Mailing List Subject: Re: Disk inconsistency X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 28 Sep 2005 11:08:18 -0000 >>>>> On Wed, 28 Sep 2005 06:25:44 -0400, >>>>> Michael Conlen said: > I have reason to believe that a set of mirrored disks became > inconsistent recently. Since reviving the disk array the system it's > attached to has become highly unstable. It appears to deadlock every > few hours. No errors, no logs, no response to keyboard, ping or other > network requests. > Each reboot takes several passes with FSCK to get the disks in to a > clean state to boot with again. > Can anyone confirm that reading data from a set of mirrored drives > which are inconsistent would cause this type of symptom? A month ago I had one drive in a raid 1 volume "intermittently fail". I started seeing occasional (as in only once or twice per week) read errors in the logs for the volume in question; however, the drive didn't fail catastrophically enough for me to identify which one of the 2 drives was bad. After this happened a few times, I started seeing exactly the behaviour that you described above. At the time, I conjectured that the mirrored disks had become inconsistent. Fortunately, I was able to identify the bad drive soon after. After replacing the drive and rebuilding the redundant data, the system has been perfectly stable. Based on my experience, I would guess that it is indeed possible that your raid 1 data is inconsistent. The difficulty is determining which of the 2 disks has good data (if either). If you have a third drive of the same model, you might try replacing in succession each of the two drives with the third and rebuilding the reduntand data. It goes without saying that backing up to tape would be a good idea before mucking about with the drives. If your system isn't stable enough to dump the raid volume to tape, try offlining one or the other of the drives to see if that helps. You didn't state your raid setup (hardware or software?). In my case, I am using hardware raid (a Mylex extremeRAID 1100 controller) with SCSI disks. Sandy