From owner-freebsd-questions@FreeBSD.ORG Tue Dec 7 18:09:57 2010 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 99C291065695 for ; Tue, 7 Dec 2010 18:09:57 +0000 (UTC) (envelope-from joost@jodocus.org) Received: from fep20.mx.upcmail.net (fep20.mx.upcmail.net [62.179.121.40]) by mx1.freebsd.org (Postfix) with ESMTP id 1018D8FC08 for ; Tue, 7 Dec 2010 18:09:56 +0000 (UTC) Received: from edge04.upcmail.net ([192.168.13.239]) by viefep20-int.chello.at (InterMail vM.8.01.02.02 201-2260-120-106-20100312) with ESMTP id <20101207180955.YJXZ1667.viefep20-int.chello.at@edge04.upcmail.net>; Tue, 7 Dec 2010 19:09:55 +0100 Received: from bps.jodocus.org ([24.132.166.223]) by edge04.upcmail.net with edge id gJ9t1f01L4pXSfL04J9uqU; Tue, 07 Dec 2010 19:09:55 +0100 X-SourceIP: 24.132.166.223 Received: from jodocus.org (localhost [IPv6:::1]) by bps.jodocus.org (8.14.2/8.14.2) with ESMTP id oB7I9q7I003093; Tue, 7 Dec 2010 19:09:52 +0100 (CET) (envelope-from joost@jodocus.org) Received: from 88.159.78.64 (SquirrelMail authenticated user joost) by jodocus.org with HTTP; Tue, 7 Dec 2010 19:09:53 +0100 Message-ID: In-Reply-To: References: Date: Tue, 7 Dec 2010 19:09:53 +0100 From: joost@jodocus.org To: "David DEMELIER" User-Agent: SquirrelMail/1.4.21 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.2.6 (bps.jodocus.org [IPv6:::1]); Tue, 07 Dec 2010 19:09:53 +0100 (CET) X-Virus-Scanned: clamav-milter 0.96.4 at bps.jodocus.org X-Virus-Status: Clean X-Cloudmark-Analysis: v=1.1 cv=kR4HTYXl3FXdRAFjS5RpcnDbDNViz/VYMWXR75RtSM0= c=1 sm=0 a=Rit7POsq_2oA:10 a=8nJEP1OIZ-IA:10 a=xqWC_Br6kY4A:10 a=hYwbykyqAAAA:8 a=TdvLcpxSfV-wO-WfGrgA:9 a=lJ9ZEb9Qe-ue7N60BtQA:7 a=gFPlNDGqlp87qAe-0_i01QXVMhEA:4 a=wPNLvfGTeEIA:10 a=VIfz433Ik-8A:10 a=3uM_M7QAeSCt-7Pm:21 a=fq-Jh841dpMBAm5Q:21 a=HpAAvcLHHh0Zw7uRqdWCyQ==:117 Cc: freebsd-questions@freebsd.org Subject: Re: can't mount root during freebsd-update 7.0 -> 8.1 on amd64 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: Tue, 07 Dec 2010 18:09:57 -0000 > 2010/12/5 : >> Hi >> >> I'm trying to upgrade a amd64 box from 7.0-RELEASE to 8.1-RELEASE with >> freebsd-update. After the first reboot the 8.1R GENERIC kernel loads >> (I'm >> using a custom kernel, so at this point the generic kernel is loaded >> manually) >> It then claims it can't mount / >> > > GEOM_PART becomes the default slicer > > Status: Committed to -CURRENT > Will appear in 8.0: sure > Author: Marcel Moolenaar & others > Web: commit message > > GEOM_PART (gpart) is a new GEOM partition class (slicer) and utility > that rolls up support for many partitioning formats (MBR, BSD, GPT > etc.) into a single code base. > > NOTE: Caveat when upgrading! GEOM_PART might interpret existing > partition tables (especially if many operating systems are present - > multi boot) differently than the previous classes. Your devices might > get renamed. > > NOTE: Some old utilities like bsdlabel may not work if the kernel > doesn't include GEOM_BSD and other old slicer classes. In other words, > bsdlabel et al don't work with GEOM_PART. > > Maybe you're encountering this trouble right now. what I would advise > you is using labels. Please do the following : > > tunefs -L root /dev/ad4a > tunefs -L var /dev/ad4e (assuming it's e on your system ?) > and for usr, tmp, etc ... > >>From a fixit environment. then edit your /etc/fstab and place > /dev/ufs/root /dev/ufs/var instead of hardcoding the device node. > Adding labels to the partitions doesn't help. I've tried both UFS and GEOM labels. You did point me in the right direction though. If I load geom_mbr my partitions show up and all is well. I think it was actually the slice that wasn't recognized. geom_part should have recognized the mbr, right? I've scanned g_part_mbr_probe() and all the checks it does should succeed. DOSMAGIC is there. The first byte of each row in the partition table is either 0x80 of 0x00. geom_part seems to do first match when probing. is there a way for me to find out which geom_part_* module is staking a claim on my mbr? I'm assuming I'm not the first person to try 8.x on amd64 with an mbr on a disk so there must be something 'special' with this system... Joost.