From owner-freebsd-hackers@FreeBSD.ORG Tue Apr 29 20:00:09 2014 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 2E3616DB for ; Tue, 29 Apr 2014 20:00:09 +0000 (UTC) Received: from louis.schedom-europe.net (louis.schedom-europe.net [193.109.184.93]) by mx1.freebsd.org (Postfix) with SMTP id 908D113D6 for ; Tue, 29 Apr 2014 20:00:08 +0000 (UTC) Received: (qmail 31099 invoked by uid 507); 29 Apr 2014 21:53:26 +0200 X-Spam-Checker-Version: SpamAssassin 3.2.5 (2008-06-10) on louis.schedom-europe.net X-Spam-Level: *********** X-Spam-Status: No, score=11.2 required=15.0 tests=BAYES_99, FH_DATE_PAST_20XX, RCVD_IN_PBL,RDNS_DYNAMIC autolearn=disabled version=3.2.5 Received: from ip-83-101-50-26.customer.schedom-europe.net (HELO webmail.malavon.com) (83.101.50.26) by louis.schedom-europe.net with SMTP; 29 Apr 2014 21:53:18 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Date: Tue, 29 Apr 2014 19:53:18 +0000 From: Benny Goemans To: freebsd-hackers@freebsd.org Subject: Intel Matrix RAID metadata messed up by geom Message-ID: <2eaf47c4d979b94ae13d6a83c7169e73@webmail.malavon.com> X-Sender: benny.goemans@belgacom.net User-Agent: Roundcube Webmail/1.0.0 X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 29 Apr 2014 20:00:09 -0000 (I originally sent this to the geom mailinglist, but I assume that list isn't meant for user questions; I've also added some more info) Hi all, First some background: I'm running 2 rather big raid (well, software raid) arrays, one RAID 0 and one RAID 5. Up to now this didn't pose much problems, aside from the occasional rebuilding for which I needed to boot my fallback windows OS. I was happy with this, but since I'm lagging behind this much in version (8- will soon disappear) I assumed that I could at least try upgrading to 10 and see what happens. So, I did: * used freebsd-update to get updates, merge and install * rebooted into BSD 10 kernel * I see messages from GEOM passing by, noticing my raid arrays * I get warnings about read-only filesystems and I realise that geom doesn't support writing to ICH RAID 5 arrays (which I read when I wanted to upgrade to 9, but seemed to have forgotten) I figure, no harm done, I'll just boot a 8.4 disk and reinstall the kernel. So I reboot ... only to find out that my drives show up as INCOMPATIBLE in the ICH configuration utility. I'm assuming that GEOM decided to overwrite critical metadata with a version that my motherboard doesn't support. I never ran any geom tools, never saw a warning in the boot which would have prompted me to abort. It just did. So, I have two questions: 1a) isn't this a bug? I'm assuming that GEOM shouldn't at least overwrite metadata without allowing the user to abort, especially not with a newer version that might not be supported anymore 1b) does this also happen when booting the live cd? if it does, it doesn't feel right 2) is there anyone out there who can help me get this meta data back to a version my mainboard supports? I really need this system up and running again (as it's my main pc) without having to reinstall everything, which would take ages. I do have backups of most critical data, but I'd still like to have the non/less-critical data as well. Some extra info: mainboard: Asus P5WDG 2 WS, with latest (official) 0805 bios ICH: 5.1.2, ICH7R I can boot the system with a BSD 10 memdisk and access the data on the BSD partitions, so it's not gone. For the ntfs partitions I'm still looking for a solution, but as I said I'd rather get the system up and running again than try to backup everything that I need and rebuild. The risk that I'm forgetting something is simply too big. What I'm currently doing/trying: * backing up as much as I can * trying to find more information of the metadata written by geom or the ICH, without much luck ** I figure the changes won't be huge between what GEOM has written and what the original data was; if I find a datasheet I might try to fix it manually * figuring out a way to restore the correct metadata without trying to understand geom source code, as this is _way_ over my skill set; it would likely take years for me to figure it out * checking out unofficial bios'es for my mainboard that may or may not have a more recent ICH on board; last resort only though as these may put me off even worse (with no more access to these disks at all) Kind regards, Benny Goemans