From owner-freebsd-current@FreeBSD.ORG Sun Jan 13 00:08:42 2013 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by hub.freebsd.org (Postfix) with ESMTP id 59466547 for ; Sun, 13 Jan 2013 00:08:42 +0000 (UTC) (envelope-from gpalmer@freebsd.org) Received: from noop.in-addr.com (mail.in-addr.com [IPv6:2001:470:8:162::1]) by mx1.freebsd.org (Postfix) with ESMTP id 31CB1EB5 for ; Sun, 13 Jan 2013 00:08:42 +0000 (UTC) Received: from gjp by noop.in-addr.com with local (Exim 4.80.1 (FreeBSD)) (envelope-from ) id 1TuB7w-000O1D-04; Sat, 12 Jan 2013 19:08:40 -0500 Date: Sat, 12 Jan 2013 19:08:39 -0500 From: Gary Palmer To: "xenophon\\+freebsd" Subject: Re: Expanding ZFS RAIDZ on the fly? Message-ID: <20130113000839.GB54865@in-addr.com> References: <50EFBAEA.3070200@zedat.fu-berlin.de> <20130111140557.GA63102@psconsult.nl> <20130111233428.GA54865@in-addr.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-SA-Exim-Connect-IP: X-SA-Exim-Mail-From: gpalmer@freebsd.org X-SA-Exim-Scanned: No (on noop.in-addr.com); SAEximRunCond expanded to false Cc: freebsd-current@freebsd.org X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 13 Jan 2013 00:08:42 -0000 On Sat, Jan 12, 2013 at 04:11:18PM -0500, xenophon\+freebsd wrote: > > -----Original Message----- > > From: owner-freebsd-current@freebsd.org [mailto:owner-freebsd- > > current@freebsd.org] On Behalf Of Gary Palmer > > Sent: Friday, January 11, 2013 6:34 PM > > > > BPR doesn't look likely, unfortunately. > > I wonder how other SANs implement restriping. The HP P4000/P4500 can > restripe logical units, and it even maintains data redundancy in the > process. Probably more relevantly, NetApp can do it in WAFL, and WAFL uses a number of concepts similar to ZFS in is design (witness the Sun/NetApp patent punch up). From what I understand of the NetApp method, they add the space to the pool and as data is written to the volume, some of it is written to the new disks. Because WAFL is a COW storage system, like ZFS, even if no new data is added, just modified in place, over time the storage is balanced out among members of the aggregate (pool). There is also a command to forcibly rebalance the data from memory. I don't know enough about ZFS to know why something similar can't be done Gary