From owner-freebsd-stable@FreeBSD.ORG Mon Apr 19 19:32:38 2010 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D288B1065670 for ; Mon, 19 Apr 2010 19:32:38 +0000 (UTC) (envelope-from petefrench@ticketswitch.com) Received: from constantine.ticketswitch.com (constantine.ticketswitch.com [IPv6:2002:57e0:1d4e:1::3]) by mx1.freebsd.org (Postfix) with ESMTP id 6CEED8FC15 for ; Mon, 19 Apr 2010 19:32:38 +0000 (UTC) Received: from dilbert.rattatosk ([10.64.50.6] helo=dilbert.ticketswitch.com) by constantine.ticketswitch.com with esmtps (TLSv1:AES256-SHA:256) (Exim 4.69 (FreeBSD)) (envelope-from ) id 1O3whE-0003mR-V7 for freebsd-stable@freebsd.org; Mon, 19 Apr 2010 20:31:52 +0100 Received: from petefrench by dilbert.ticketswitch.com with local (Exim 4.71 (FreeBSD)) (envelope-from ) id 1O3whE-000GSB-UC for freebsd-stable@freebsd.org; Mon, 19 Apr 2010 20:31:52 +0100 Date: Mon, 19 Apr 2010 20:31:52 +0100 Message-Id: To: freebsd-stable@freebsd.org From: Pete French Subject: Problems with gmirror locking up on recent 8.0-STABLE/i386 X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 19 Apr 2010 19:32:38 -0000 I have a machine used as a firewall which boots from a pair of drives mirrored using gmirror. This has been in situ for a long time, and was upgraded to 8.0 a while ago. Just before the weekend I upgraded the kernel to the latest STABLE to get the new em code. After doing this I observed that at seemingly random points the disc system would 'lockup' by which I mean that all networking continued fine, but if you tried to do anything which would read or write the disc then that would pause forever. Only fixed by reboot, and on reboot the array was degraded, and always resynced the same drive. "dying drive" I thought to myself - seemed a resonable conclusion. Except today I took another machine (completelt different hardware BTW) and build a new system with different drives, but the same gmirrored configuration - and an hour after I had installed it, it locked up in exactly the same way. Just after I upgraded it to match the version of the kernel running onthe original one in fact. So now I am wondering if there is some unfortunate problem which has crept into gmirror in the last few weeks. Has anyone else seen anything like this ? -pete.