From owner-freebsd-current Sat Jan 31 17:42:41 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id RAA20481 for current-outgoing; Sat, 31 Jan 1998 17:42:41 -0800 (PST) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from allegro.lemis.com (allegro.lemis.com [192.109.197.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA20465 for ; Sat, 31 Jan 1998 17:42:37 -0800 (PST) (envelope-from grog@lemis.com) Received: from freebie.lemis.com (freebie.lemis.com [192.109.197.137]) by allegro.lemis.com (8.8.7/8.8.5) with ESMTP id MAA29489; Sun, 1 Feb 1998 12:07:51 +1030 (CST) Received: (from grog@localhost) by freebie.lemis.com (8.8.8/8.8.7) id MAA07909; Sun, 1 Feb 1998 12:07:51 +1030 (CST) (envelope-from grog) Message-ID: <19980201120750.49180@lemis.com> Date: Sun, 1 Feb 1998 12:07:50 +1030 From: Greg Lehey To: Karl Denninger Cc: Brian Tao , current@FreeBSD.ORG Subject: Re: RAID controllers - folks, check this thing out References: <19980131155527.19192@mcs.net> <19980131185454.07579@mcs.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.84e In-Reply-To: <19980131185454.07579@mcs.net>; from Karl Denninger on Sat, Jan 31, 1998 at 06:54:54PM -0600 Organization: LEMIS, PO Box 460, Echunga SA 5153, Australia Phone: +61-8-8388-8286 Fax: +61-8-8388-8725 Mobile: +61-41-739-7062 WWW-Home-Page: http://www.lemis.com/~grog Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG X-To-Unsubscribe: mail to majordomo@FreeBSD.org "unsubscribe current" On Sat, Jan 31, 1998 at 06:54:54PM -0600, Karl Denninger wrote: > On Sat, Jan 31, 1998 at 05:25:45PM -0500, Brian Tao wrote: >> On Sat, 31 Jan 1998, Karl Denninger wrote: >>> >>> RAID 5, due to the way it stripes parity across the volumes, has a >>> "sweet spot" in performance at 5 spindles. >> >> What "way" is that? On a given stripe, one drive provides the >> parity block, the choice of drive staggered across consecutive >> stripes. There may be an issue with small, sequential writes on a >> RAID 5 set with a large number of drives, but I can't think of any >> reason why five drives should be magical. > > A single write which does not consume an entire stripe requires that the > entire stripe be READ FIRST in order to recompute parity. If you're using ufs, all writes to the device will be ufs blocks. There is obviously an advantage in laying out your stripes accordingly. This means you would need 2**n+1 drives, in practice either 3, 5, or 9. Of course, if you're only writing extremely infrequently, you may find other advantages to a different number of spindles which cancel out this one. Greg