From owner-freebsd-questions@FreeBSD.ORG Fri Jul 31 20:10:41 2009 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 15BD91065678 for ; Fri, 31 Jul 2009 20:10:41 +0000 (UTC) (envelope-from af.gourmet@videotron.ca) Received: from relais.videotron.ca (relais.videotron.ca [24.201.245.36]) by mx1.freebsd.org (Postfix) with ESMTP id DE3D38FC1B for ; Fri, 31 Jul 2009 20:10:40 +0000 (UTC) (envelope-from af.gourmet@videotron.ca) MIME-version: 1.0 Content-transfer-encoding: 8BIT Content-type: text/plain; charset=UTF-8 Received: from [192.168.0.51] ([96.21.103.185]) by VL-MO-MR003.ip.videotron.ca (Sun Java(tm) System Messaging Server 6.3-4.01 (built Aug 3 2007; 32bit)) with ESMTP id <0KNN006PSXDRMNB0@VL-MO-MR003.ip.videotron.ca> for freebsd-questions@freebsd.org; Fri, 31 Jul 2009 16:10:40 -0400 (EDT) Message-id: <4A734FBF.6040308@videotron.ca> Date: Fri, 31 Jul 2009 16:10:39 -0400 From: PJ User-Agent: Thunderbird 2.0.0.22 (Windows/20090605) To: Roland Smith References: <4A71DB2A.4040401@videotron.ca> <20090730190458.GA36265@slackbox.xs4all.nl> <4A71F297.1080903@videotron.ca> <20090730220618.GA40281@slackbox.xs4all.nl> <4A734215.6080003@videotron.ca> <20090731195121.GA92292@slackbox.xs4all.nl> In-reply-to: <20090731195121.GA92292@slackbox.xs4all.nl> Cc: freebsd-questions@freebsd.org Subject: Re: how to boot or access problem file system 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: Fri, 31 Jul 2009 20:10:41 -0000 Roland Smith wrote: > On Fri, Jul 31, 2009 at 03:12:21PM -0400, PJ wrote: > >> Roland Smith wrote: >> >>> On Thu, Jul 30, 2009 at 03:20:55PM -0400, PJ wrote: >>> >>> >>>> Roland Smith wrote: >>>> >>>> >>>>> On Thu, Jul 30, 2009 at 01:40:58PM -0400, PJ wrote: >>>>> >>>>> >>>>> >>>>>> What can be done to access a file system that seems to have the boot >>>>>> sector screwed up? >>>>>> >>>>>> >>> I forgot to mention that your boot sector is fine. If it were screwed >>> up, you wouldn't get to the boot prompt. >>> >>> Since the boot code cannot locate your kernel, there are several things >>> that could have gone wrong. See below. >>> >>> >>> >>> >>>>>> The /usr files should be ok but how to access? >>>>>> >>>>>> >>>>> Use fsck_ffs to try and repair the filesystem. >>>>> >>>>> >>>>> >>>> how can I use it if I can't boot or access the file system? >>>> >>>> >>> Use a livefs cd or use the Fixit option in the main menu of sysinstall >>> on an install disk. That should get you a shell where you can run >>> fsck_ffs on your disk partitions. >>> >>> If you have booted from CD, list the disk devices with e.g. 'ls >>> /dev/ad*'. If you have SCSI drives, use 'da' instead of 'ad'. >>> What does that command list? On my machine, I'll get >>> something like this: >>> >>> /dev/ad4 /dev/ad4s1d /dev/ad6 /dev/ad6s1d >>> /dev/ad4s1 /dev/ad4s1e /dev/ad6s1 /dev/ad6s1e >>> /dev/ad4s1a /dev/ad4s1f /dev/ad6s1a /dev/ad6s1f >>> /dev/ad4s1b /dev/ad4s1g /dev/ad6s1b /dev/ad6s1g >>> /dev/ad4s1c /dev/ad4s1g.eli /dev/ad6s1c /dev/ad6s1g.eli >>> >>> If you only see e.g. /dev/ad4 and /dev/ad6, your slice table has been >>> overwritten (with fdisk) and your data is lost. If you see /dev/ad4s1 >>> but not /dev/ad4s1a-g, the BSD partitions have been removed and your >>> data is lost as well. >>> >>> Since there is only one slice on both ad4 and ad6 (otherwise you'd see >>> /dev/ad4s2x) The next step is to examine the disk labels: >>> >>> bsdlabel /dev/ad4s1 >>> # /dev/ad4s1: >>> 8 partitions: >>> # size offset fstype [fsize bsize bps/cpg] >>> a: 1024000 16 4.2BSD 2048 16384 64008 >>> b: 16777216 1024016 swap >>> c: 976768002 0 unused 0 0 # "raw" part, don't edit >>> d: 4194304 17801232 4.2BSD 2048 16384 28528 >>> e: 104857600 21995536 4.2BSD 2048 16384 28528 >>> f: 41943040 126853136 4.2BSD 2048 16384 28528 >>> g: 807971826 168796176 4.2BSD 2048 16384 0 >>> >>> This tells us that the a, d, e, f and g partition are carrying a BSD >>> filesystem, and should be checked with fsck_ffs. >>> >>> Try these steps and report back what you find. >>> >>> >>> >>>>>> I don't have a problem with irrecoverable files, I would just finally >>>>>> understand how things work and what can be done on FBSD. >>>>>> >>>>>> >>>>> Make regular backups. Especially before big upgrades. >>>>> >>>>> >>>>> >>>> Maybe the real problem is that the manual is too screwed up (why are >>>> there so many problems being brought up on the mailing lists? we can't >>>> all be that stupid.) >>>> >>>> >>> It is a mailing list for questions. Ipso facto you'll see questions and >>> problems on this list. People who are not having problems will not be >>> posting very much. :-) >>> >>> As to the handbook, this is by necessity written by people who are >>> knowledgeable on the subject they write on. Unfortunately this sometimes >>> lead to really basic steps/assumptions being skipped because they are >>> self-evident for the writer. If you gain enough knowledge about a >>> subject it becomes really hard to write for people new to the system >>> because you've internalized a lot of stuff by then. >>> >>> If you have specific questions about parts of the handbook, ask. >>> >>> >> I get the impression that my disks have all been overwritten; it's >> > > Don't have impressions. Get the data. Boot from a livefs CD and start a > shell as explained in in some of my previous messages. Then use the commands > listed above to check your filesystems. *And report back wat you found*. > > >> rather strange that in the instructions to upgrade it says to not change >> anything on the Newfs... and that files would not be overwritten... is >> that at fact? >> > > What instructions are you referring to? Neither the handbook section > [http://www.freebsd.org/doc/en/books/handbook/updating-freebsdupdate.html] > nor the manual page for freebsd-update mention newfs at all! Nor should they. > > Well, it sounds like we should be ok, but the instructions I am speaking of are those on the installation disks ... when one selects the update to a newer version and then as asked to set the slice names /; swap; /tmp; and /var. If I have made an error in this, then I can only put the full blame on whoever created the installation system... there should be the most obvious checks & balances about such an installation/upgrade as - warnings about what the installation is going to do, and are you sure this is what you want to do; and warnings that any existing files will be overwritten... I have always cringed a bit when installing as there were never any decent confirmations as to what was happening... even Microsoft is pretty good with warning. So, as I mentioned, I cannot access any disks as "Permission denied" This seems to be a dead end. :-( >> If that is true, then surely it should be possible to recover files in >> the /usr /var and /tmp directories. If the disks have not been >> overwritten... I think there was a huge misinformation gap here if this >> is not so... >> > > For an upgrade, the filesystems are not overwritten. Only a new install > creates new filesystems. > > Please boot from a livefs CD and check the filesystems on the harddisk > as explained before and report the results. > > Roland > -- Hervé Kempf: "Pour sauver la planète, sortez du capitalisme." ------------------------------------------------------------- Phil Jourdan --- pj@ptahhotep.com http://www.ptahhotep.com http://www.chiccantine.com/andypantry.php