From owner-freebsd-questions Sun May 27 8:50:52 2001 Delivered-To: freebsd-questions@freebsd.org Received: from mail2.mediadesign.nl (md2.mediadesign.nl [212.19.205.67]) by hub.freebsd.org (Postfix) with SMTP id 8098737B422 for ; Sun, 27 May 2001 08:50:48 -0700 (PDT) (envelope-from alson@mediadesign.nl) Received: (qmail 1678 invoked by uid 1002); 27 May 2001 15:50:47 -0000 Date: Sun, 27 May 2001 17:50:47 +0200 From: Alson van der Meulen To: freebsd-questions@freebsd.org Subject: Re: disaster recovery - killed the partition table on my bootdisk Message-ID: <20010527175047.G13125@md2.mediadesign.nl> Mail-Followup-To: freebsd-questions@freebsd.org References: <20010527162223.Z26314@xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.15i In-Reply-To: <20010527162223.Z26314@xs4all.nl>; from rene@xs4all.nl on Sun, May 27, 2001 at 04:22:24PM +0200 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Sun, May 27, 2001 at 04:22:24PM +0200, rene@xs4all.nl wrote: > I've been a very stupid boy today. > > I was trying to newfs one of my 9G disks (ad0) and hang it in /etc/fstab. > When I applied the 'manual' handbook procedure to this disc, I made a > critical error; > > # dd if=/dev/zero of=/dev/rda1 bs=1k count=1 > # disklabel -Brw da1 auto > # disklabel -e da1 # create the `e' partition > # newfs -d0 /dev/rda1e > # mkdir -p /1 > # vi /etc/fstab # add an entry for /dev/da1e > # mount /1 > > when I ran commands 1-3 on my da0 (boot)disk! > > It gave me a warning error on the second disklabel command, which kinda > made me go 'whoops; shouldnt be doing this on this disc'. I then (scared) > proceeded to use /stand/sysinstall to newfs the ad0 disk. > > When I rebooted, however, my system wouldn't come online anymore. (yeah, > yeah, surprise, surprise ;)..... unhapiness. > > I have grabbed my original 4.2-Release bootCD, and am able to use that to > view the installation menu. When viewing the disklabel for my da0 disk, I > see that there are no partitions and/or slices. > > My question is very simple; can I still resque the data on my da0 disk, > and if so, how? I know the approximate layout of my disk; it was > > 300M for / > 1500M for /usr > 200M for /var > > but ofcourse, I am probably forgetting how much /swap I had ;((( have a look at gpart, it can guess your partition table, dunno if it can reconstruct disklabels. i used it once to reconstruct the primary partition table of a disk. i'm afraid it can't reconstruct your disklabel though... you can at least try it. to reconstruct it, it's most easy to but the hd in another box that can boot some *bsd thing (or some other os that can run gpart, like linux). if you don't have another box, you could try to install freebsd from scratch on a different disk. gpart is in the ports collection. you might also wanna look at ffsrecover (in ports) for recovering stuff from ufs partittions. you might be able to reconstruct your disklabel by creating an entry for the first fs, spanning the whole label. if you mount it, you might be able to see the size using df. then try resizing it around that size, 'till you find the second fs. this will be difficult with a swap partition. if your data is plain ascii, just open it in vi in read-only mode (vi /dev/ad0c), and search for the stuff you want, then save it to a file on an other disk. this only works for small amounts of plain ascii data though... maybe there are some tools around that can search for ufs superblocks, gpart might be able to do it... Alson To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message