From mboxrd@z Thu Jan 1 00:00:00 1970 From: Heinz Mauelshagen Subject: Re: [PATCH] Add metadata writing and hot spare processing to hostraid driver in dmraid Date: Sat, 4 Mar 2006 10:10:49 +0100 Message-ID: <20060304091049.GC2966@redhat.com> References: <4408E5F0.50208@us.ibm.com> Reply-To: mauelshagen@redhat.com, device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <4408E5F0.50208@us.ibm.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: "ATARAID (eg, Promise Fasttrak, Highpoint 370) related discussions" Cc: dm-devel@redhat.com, Chris McDermott List-Id: dm-devel.ids On Fri, Mar 03, 2006 at 04:57:20PM -0800, Darrick J. Wong wrote: > Hi all, Hi Darrick. > > Now that dmraid supports Adaptec HostRAID, I've turned my attention to > enabling support for writing metadata back to the disks, and > implementing rudimentary spare handling in RAID1 configurations. > > The attached patch (which is admittedly very rough) does two key things. Will look into it on Mon. > First, it implements asr_write in a more meaningful manner than what's > in rc10 right now; with that part of the patch, dmraid gains the ability > to reconfigure an array while the system is running and save the changes > out to disk. To do this, I had to change a struct to a pointer to a > struct in asr_reservedblock to to get write_metadata to work properly. > Also, it turns out that a few more things in asr_write had to happen > (and in the correct order) for writes to succeed. If we fail to write > the metadata correctly, the HostRAID BIOS wipes the entire sector during > the next reboot. > > The second part of the patch stores spare devices in a dummy .asr_spares > array. When an I/O error is noticed (via event_io), a suitable spare > drive is pulled from the spare pool and inserted into the failing array. > I haven't had a chance to test this code rigorously yet, because the > dmeventd portions of dmraid aren't implemented yet. > > So, there are two things that I'd like to see added: online > reconfiguration support in dmraid (which I can now test via asr_write) > and online spare handling. Heinz, are you working on either of these > features, or should I start writing them? I am into device-mapper core work right now so please, go ahead :) Heinz > > --D > _______________________________________________ > Ataraid-list mailing list > Ataraid-list@redhat.com > https://www.redhat.com/mailman/listinfo/ataraid-list =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Heinz Mauelshagen Red Hat GmbH Consulting Development Engineer Am Sonnenhang 11 Cluster and Storage Development 56242 Marienrachdorf Germany Mauelshagen@RedHat.com +49 2626 141200 FAX 924446 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-