From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Bryn M. Reeves" Subject: Re: partitioning multipath device Date: Mon, 25 Jan 2010 12:32:25 +0000 Message-ID: <1264422745.8461.5559.camel@localhost> References: <20100122192935.GC31538@mtholyoke.edu> Reply-To: device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20100122192935.GC31538@mtholyoke.edu> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: device-mapper development List-Id: dm-devel.ids On Fri, 2010-01-22 at 14:29 -0500, Ron Peterson wrote: > The multipath-tools FAQ has the following to say about partitioning > multipath devices. > > >From http://christophe.varoqui.free.fr/faq.html > > > 6. How do I partition my multipathed devices > > Partition a constituant path instead. And reload the map through multipath(8) > > I take this to mean, for example, that one should partition /dev/sdX, > rather than the corresponding coalesced /dev/mapper/... device, correct? > > My question is: why? Paritioning the /dev/mapper device appears to > work. Is there an operational difference? I've always done this > exactly as advised, but I know someone who didn't, and we're wondering > if it matters. It doesn't actually matter too much. If you're partitioning before assembling maps then you may as well do it on a path. The common partitioning tools should do the right thing if you point them at the multipath device too though (although you may need to arrange for kpartx to get run again, or run it manually to refresh). Regards, Bryn.