From owner-cvs-all Fri Oct 26 20:50:56 2001 Delivered-To: cvs-all@freebsd.org Received: from wantadilla.lemis.com (wantadilla.lemis.com [192.109.197.80]) by hub.freebsd.org (Postfix) with ESMTP id 3BCF937B40F; Fri, 26 Oct 2001 20:49:41 -0700 (PDT) Received: by wantadilla.lemis.com (Postfix, from userid 1004) id C55646ACCD; Sat, 27 Oct 2001 12:37:50 +0930 (CST) Date: Sat, 27 Oct 2001 12:37:50 +0930 From: Greg Lehey To: Matthew Jacob Cc: Doug Rabson , Luigi Rizzo , John Baldwin , Jonathan Lemon , cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org Subject: Re: RAID-5 parity calculations (was: cvs commit: src/sys/dev/fxp if_fx) Message-ID: <20011027123750.D8085@wantadilla.lemis.com> References: <20011027111145.A7846@wantadilla.lemis.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit User-Agent: Mutt/1.2.5i In-Reply-To: ; from mjacob@feral.com on Fri, Oct 26, 2001 at 07:46:27PM -0700 Organization: The FreeBSD Project Phone: +61-8-8388-8286 Fax: +61-8-8388-8725 Mobile: +61-418-838-708 WWW-Home-Page: http://www.FreeBSD.org/ X-PGP-Fingerprint: 6B 7B C3 8C 61 CD 54 AF 13 24 52 F8 6D A4 95 EF Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Friday, 26 October 2001 at 19:46:27 -0700, Matt Jacob wrote: >>> I guess the real question is: 'can you get the parity calculations >>> done in time so that the entire stripe can go out together'. >> >> Why? > > Because it's best if you can push stripes in some configurations out > together- essentially a software spindle sync. I don't see any advantage in this. Spindle sync is good for RAID-3, but nobody I know uses that on general-purpose machines. It also has the bad disadvantage that it implies that you always transfer complete bands, which is bad both for throughput and for memory requirements. On Friday, 26 October 2001 at 19:49:26 -0700, Matt Jacob wrote: >>> Instantaneously? Well, how long does it take to checksum 6 kB (the >>> average request size)? That's 1500 words, say 10000 instructions. At >>> 1 GHz, that's 10 µs, which is completely negligible compared to the >>> minimum four I/O transactions needed, which add up to about 25 ms. >>> > > BTW- for the Convex in question, the average request size was 64K. Yes, that's the "always transfer whole bands" syndrome. It makes people choose artificially small stripes, but it does simplify the calculations. To get an idea how horrendous they are, look at vinumraid5.c, specifically bre5(). Greg -- See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message