From mboxrd@z Thu Jan 1 00:00:00 1970 From: Oleg Drokin Subject: Re: LVM array, one disk failed Date: Sat, 28 Dec 2002 11:39:42 +0300 Message-ID: <20021228113942.A6599@namesys.com> References: <1041044851.5964.32.camel@ganga> Mime-Version: 1.0 Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com Content-Disposition: inline In-Reply-To: <1041044851.5964.32.camel@ganga> List-Id: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Tuomas Jormola Cc: reiserfs-list@namesys.com Hello! On Sat, Dec 28, 2002 at 05:07:31AM +0200, Tuomas Jormola wrote: > I have a 110GiB LVM array constructed using EVMS. It consists of three > IDE hard disks from which the last one died the other day. The > filesystem is ReiserFS, of course. Array is a linear one and there was > about 15GiB free space when disk died so I'm hoping there was a somewhat > resonable way to recover the data on two disks that are full and in > front of the array. The failed disk is totally dead, not just a couple > of bad sectors. It spins but it's not detected on boot. When I try to > mount the volume with only two disks connected, it get this: > sh-2029: reiserfs read_bitmaps: bitmap block (#19234816) reading failed > reiserfs_read_super: unable to read bitmap This is hardware problem ;) But if your raid setup does not use any stripping, this is possible to rercover from such a situation. The easiest way that comes to mind is to add another disk instead failed one so that all of the FS blocks still accessible. Then you can mount the FS and all the data would be intact (only bitmaps and other stuff from failed disk would be corrupted). If you do not plan to add anymore drives or if you just don't have spare drive, you probably can run reiserfsck --rebuild-sb and after couple of questions it will write updated superblock that respects changed filesystem size (actually for this to work you first need to tell your LVM to report current volume size that is now shorter than it was at creation time). Bye, Oleg