From owner-freebsd-scsi@FreeBSD.ORG Thu Jul 10 09:23:00 2014 Return-Path: Delivered-To: freebsd-scsi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 957F7570 for ; Thu, 10 Jul 2014 09:23:00 +0000 (UTC) Received: from gw.zefyris.com (sabik.zefyris.com [IPv6:2001:7a8:3c67:2::254]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 273A7252F for ; Thu, 10 Jul 2014 09:22:59 +0000 (UTC) Received: from sekishi.zefyris.com (sekishi.zefyris.com [IPv6:2001:7a8:3c67:2::12]) by gw.zefyris.com (8.14.5/8.14.5) with ESMTP id s6A9Mprk008427; Thu, 10 Jul 2014 11:22:51 +0200 (CEST) Date: Thu, 10 Jul 2014 11:22:51 +0200 From: Francois Tigeot To: FreeBSD-scsi Subject: Data corruption with the mfi(4) driver Message-ID: <20140710092251.GA1206@sekishi.zefyris.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.23 (2014-03-12) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.4.3 (gw.zefyris.com [IPv6:2001:7a8:3c67:2::254]); Thu, 10 Jul 2014 11:22:52 +0200 (CEST) X-BeenThere: freebsd-scsi@freebsd.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: SCSI subsystem List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 10 Jul 2014 09:23:00 -0000 Hi, I have encountered catastrophic data corruption with LSI RAID adapters from recent Dell servers (R720xd). The adapters apparently worked fine at first but eventually destroyed the filesystems of volumes experiencing a high load. I found out the hard way a Postgres database import was a sure way to reproduce the problem. Switching to the mrsas(4) driver made my test machine stable again; I have yet to encounter a similar issue with it. Now, this particular server was running under DragonFly and not FreeBSD but the mfi(4) driver is mostly the same under both operating systems and I have found reports of data corruption issues with mfi(4) and recent LSI adapters in the archives of this list. DragonFly will most likely switch to use mrsas(4) by default for at least the LSI Thunderbolt serie of adapters to avoid data loss. This mail from the mfi(4) and mrsas(4) maintainer contains more details, including a partial list of impacted adapters: http://lists.dragonflybsd.org/pipermail/users/2014-July/128703.html -- Francois Tigeot