From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com (mx1.redhat.com [172.16.48.31]) by int-mx1.corp.redhat.com (8.13.1/8.13.1) with ESMTP id n6R6JYbb020276 for ; Mon, 27 Jul 2009 02:19:34 -0400 Received: from eastrmmtao105.cox.net (eastrmmtao105.cox.net [68.230.240.47]) by mx1.redhat.com (8.13.8/8.13.8) with ESMTP id n6R6JIPL028990 for ; Mon, 27 Jul 2009 02:19:18 -0400 Received: from eastrmimpo03.cox.net ([68.1.16.126]) by eastrmmtao105.cox.net (InterMail vM.7.08.02.01 201-2186-121-102-20070209) with ESMTP id <20090727061918.WPID14338.eastrmmtao105.cox.net@eastrmimpo03.cox.net> for ; Mon, 27 Jul 2009 02:19:18 -0400 Message-ID: <4A6D46E5.2000100@cox.net> Date: Mon, 27 Jul 2009 01:19:17 -0500 From: Ron Johnson MIME-Version: 1.0 Subject: Re: [linux-lvm] lvm2 confused about double UUID References: <4A6CF9C0.6020902@cox.net> <4A6D1739.2040708@cox.net> In-Reply-To: Content-Transfer-Encoding: 7bit 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"; format="flowed" To: linux-lvm@redhat.com On 2009-07-26 22:30, Christian Kujau wrote: > On Sun, 26 Jul 2009 at 21:55, Ron Johnson wrote: >> So the issue seems really the deeper fact that you've got two physical devices >> with the same UUID. All your lvm issues tumble out from there. >> >> Is that a good summary? > > Yes, that's it. I know, having 2 physical volumes with the same UUID is > not cool and is certainly not supported, but due to the split that's what > I have here now and I assumed that I could convince LVM via the filter > directive to use a certain device for vg02. I mean, vg02 is working fine, > but I really want to format either sdb or sdc, but I'm afraid of > destroying something in the LVM layout... > > Let me ask another thing: when I assume for a moment that "pvs" is lying > and really sdc is used for vg02, I'd go ahead and 1) pvcreate sdb, thus > assigning a new UUID to sdb and 2) remove the filter from lvm.conf. LVM > should not complain any more (since we have 2 different UUIDs now) and I/O > still goes to sdc, as it is now apparently. If things go wrong, how can I > restore the LVM header on sdb? (Would "pvcreate --uuid " do the > trick? Or restoring with dd? How many bytes are changed when using > pvcreate?) Now you're getting beyond my pay grade. But with uuidgen and tune2fs, you could give sdc1 a new UUID. That would hose lvm, though, and you'd have to do surgery in /etc/lvm. I'd go the safe/sure route and backup sdc2 if possible (have you written anything new to it since breaking the mirror?), and start over, dropping it from lvm. refdisking it (is there such a word?), etc, etc. -- Scooty Puff, Sr The Doom-Bringer