From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx2.redhat.com (mx2.redhat.com [10.255.15.25]) by int-mx2.corp.redhat.com (8.13.1/8.13.1) with ESMTP id l5L7htgO023537 for ; Thu, 21 Jun 2007 03:43:56 -0400 Received: from carpentier.vpsland.com (carpentier.vpsland.com [207.210.113.97]) by mx2.redhat.com (8.13.1/8.13.1) with ESMTP id l5L7hs7M026704 for ; Thu, 21 Jun 2007 03:43:54 -0400 Received: from localhost (localhost.localdomain [127.0.0.1]) by carpentier.vpsland.com (Postfix) with ESMTP id 670131822A for ; Thu, 21 Jun 2007 09:43:52 +0200 (CEST) Received: from carpentier.vpsland.com ([127.0.0.1]) by localhost (carpentier.vpsland.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xiqtSHVvWvyz for ; Thu, 21 Jun 2007 09:43:34 +0200 (CEST) Received: from soetens.vdberg.org (h8441201224.dsl.speedlinq.nl [84.41.201.224]) by carpentier.vpsland.com (Postfix) with ESMTP id 8E3B218228 for ; Thu, 21 Jun 2007 09:43:32 +0200 (CEST) Received: from [192.168.0.33] (82-170-10-11.dsl.ip.tiscali.nl [82.170.10.11]) by soetens.vdberg.org (Postfix) with ESMTP id 5A26A14C04F for ; Thu, 21 Jun 2007 09:43:32 +0200 (CEST) Message-ID: <467A2C22.6030503@vdberg.org> Date: Thu, 21 Jun 2007 09:43:30 +0200 From: Richard van den Berg MIME-Version: 1.0 Subject: Re: [linux-lvm] Does pv failure effect whole vg? References: 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" To: LVM general discussion and development Stuart D. Gathman wrote: > If the system truly won't boot with a failed disk, that kind of adds another > reason why current LVM mirroring support is useless. > > If you build your VG on raid (md devices or hardware), you solve > the problem for now. > That is what I am doing. The test I did was: lv00 and lv01 are part of vg00 which consists of /dev/md2 and /dev/sdb3 When booting from cdrom (systemrescuecd) md2 failed to come up, so vg00 did not come up. I then manually brought up md2 and split the vg, and added sda3 in the mix: lv00 is part of vg00 which consists of /dev/md2 lv01 is part of vg01 which consists of /dev/sdb3 and /dev/sda3 At the next reboot from cdrom vg01 came up fine, vg00 did not. At least after the split if sda or sdb fails, md2 should still come up and so should vg00, but I will loose vg01. Keeping all lvs in vg00 would have given me an unbootable system after 1 disk failure. Sincerely, Richard van den Berg PS: I still think it sucks that I could not split the vg online.