From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jonathan Brassow Subject: Re: [PATCH v3 0/8] dm-raid (raid456) target Date: Mon, 10 Jan 2011 14:14:59 -0600 Message-ID: <5F7256B3-BE4F-489F-8B2B-0E830A8BBE0B@redhat.com> References: <1292899065-31943-1-git-send-email-snitzer@redhat.com> <20101221141443.5aea0b12@notabene.brown> <20110105223628.GA13164@redhat.com> <20110106214645.2bd321f4@notabene.brown> <20110110113718.1e042548@notabene.brown> Mime-Version: 1.0 (Apple Message framework v936) Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20110110113718.1e042548@notabene.brown> Sender: linux-raid-owner@vger.kernel.org To: NeilBrown Cc: Mike Snitzer , agk@redhat.com, dm-devel@redhat.com, linux-raid@vger.kernel.org List-Id: linux-raid.ids On Jan 9, 2011, at 6:37 PM, NeilBrown wrote: >>> Also: >>> >>> 3: <#raid_devs> .. >>> >>> This doesn't allow offsets into the various devices like dm-stripe, >>> dm-linear, dm-raid1 all do. >> >> This is a conscious choice. The offset was discarded because it is >> unneeded. > > Famous last words!!! :) > > When reshaping a RAID5/6 to change the chunk size or layout, or when > converting an N drive RAID5 to N+1 drive RAID6, it makes it a lot > easier if > you can can change the data offset at the same time. Then you don't > have to > keep backing up data. md cannot do this yet but I hope to add the > code soon > and because 'data_offset' is a distinct value it is conceptually > easy to do. > If 'data_offset' is assumed to be zero, then it isn't. I guess I don't understand this part. Why would you need to change the data offset? Especially the offset as it relates to the individual drives? > But I don't know to what extent you are planning on supporting > reshaping of > arrays.... I don't see a reason why this can't be completely supported... brassow