From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Wed, 11 Jul 2001 10:15:13 -0500 From: Steven Lembark Subject: Re: [linux-lvm] Question about LVM Message-ID: <237260000.994864512@dizzy> In-Reply-To: References: MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: linux-lvm-admin@sistina.com Errors-To: linux-lvm-admin@sistina.com Reply-To: linux-lvm@sistina.com List-Help: List-Post: List-Subscribe: , List-Unsubscribe: , List-Archive: List-Id: Content-Type: text/plain; charset="us-ascii" To: linux-lvm@sistina.com - fox@theendoftime.net on 07/11/01 06:43:09 -0600: > Hi; > > I ran into this issue with AIX lvm. A hard drive died and the IBM > hardware monkeys removed the drive before mirroring could be broken > (removing lv copies down to one, and reducing the volume group by the one > dead drive) > > However, because the drive had been removed before this could be done, the > VGDA still had the old PVID of the old hdisk and would not let me remove > it by the cannonical hdisk name. The eventual solution was to run the > rmlvcopy command and reducevg command with the PVID instead of the > cannonical hdisk name. There is no such command as rmlvcopy (or anything > related as far as I can tell with Linux) but the reducevg is the same as > vgreduce. I am wondering if it's possible for a PV to be removed from the > VG in such a way that the PV path (device name?) no longer exists in the > system and thus, you wouldn't be able to do a vgreduce with the PV path > (device name?) In this case, is it possible to run the vgreduce command > with the PVID of the removed drive? Or does the Linux LVM handle this > automagically? Is this even a possible case under Linux? I have to admit > I'm still new to the Linux LVM after having worked with AIX LVM for years > now. Usual fix for this is to export the VG and re-import it w/ the required PV's. At that point your VG comes back with a legit set of disks. Then you can vgextent onto the new device. sl