From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com (ext-mx03.extmail.prod.ext.phx2.redhat.com [10.5.110.7]) by int-mx04.intmail.prod.int.phx2.redhat.com (8.13.8/8.13.8) with ESMTP id o0BH8wBD021245 for ; Mon, 11 Jan 2010 12:08:59 -0500 Received: from nerv.dolda2000.com (1-1-3-7a.rny.sth.bostream.se [82.182.133.20]) by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id o0BH8ffw003364 for ; Mon, 11 Jan 2010 12:08:42 -0500 Received: from [IPv6:2002:52b6:8514:200:21d:7dff:fea1:197] ([IPv6:2002:52b6:8514:200:21d:7dff:fea1:197]) (authenticated bits=0) by nerv.dolda2000.com (8.14.3/8.14.3/Debian-5) with ESMTP id o0BGsIlt017754 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Mon, 11 Jan 2010 17:54:19 +0100 From: Fredrik Tolf Date: Mon, 11 Jan 2010 17:54:18 +0100 Message-ID: <1263228858.2845.12.camel@pc7.dolda2000.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Subject: [linux-lvm] Removing a PV from a mirror Reply-To: LVM general discussion and development List-Id: LVM general discussion and development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , List-Id: Content-Type: text/plain; charset="us-ascii" To: linux-lvm@redhat.com Hi list, I've got a mirror volume with two S-ATA disks, one of which started getting bad sectors recently, and I want to replace it. I'm having trouble removing it from the volume, however. When googling around on the subject, I've read everywhere that I shouldn't have to do anymore than remove it, and LVM should discover that and reconvert the mirror back into a linear volume. To be a bit more careful, though, I ran "echo 1 >/sys/block/sdh/device/delete" to have it removed properly in software first. Much to my surprise, shall we say, that just made the LV die a horrible death; it started returning EIO for every access to it, and I had to reboot the server to get it working again. Again, this is the proper procedure according to what I've been reading on the net; did I do it wrong, somehow, or are my sources completely off? (And if this doesn't work, then what would actually happen if the disk were to fail physically?) As I rebooted the server, I tried disconnecting the disk in advance, but that made LVM unable to bring the VG back online at all. How am I supposed to take this disk out of the volume? If I should use `lvconvert -m 0 $LV', then how do I specify which disk is to be removed? I'm using a stock Debian stable (Lenny) installation, meaning Linux 2.6.24 and LVM2 2.02.39. Appreciative for any answers, Fredrik Tolf