From owner-freebsd-hackers Thu Jun 15 23:31:33 2000 Delivered-To: freebsd-hackers@freebsd.org Received: from sydney.worldwide.lemis.com (mg131-243.ricochet.net [204.179.131.243]) by hub.freebsd.org (Postfix) with ESMTP id 91DD737BD52 for ; Thu, 15 Jun 2000 23:31:24 -0700 (PDT) (envelope-from grog@sydney.worldwide.lemis.com) Received: (from grog@localhost) by sydney.worldwide.lemis.com (8.9.3/8.9.3) id EAA06568; Wed, 14 Jun 2000 04:26:23 -0700 (PDT) (envelope-from grog) Date: Wed, 14 Jun 2000 04:26:23 -0700 From: Greg Lehey To: Jaye Mathisen Cc: hackers@FreeBSD.ORG Subject: Re: Has anybody looked at RAID-5 vinum performance on the "whole system"? Message-ID: <20000614042623.N5769@sydney.worldwide.lemis.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0pre2i In-Reply-To: Organization: LEMIS, PO Box 460, Echunga SA 5153, Australia Phone: +61-8-8388-8286 Fax: +61-8-8388-8725 Mobile: +61-418-838-708 WWW-Home-Page: http://www.lemis.com/~grog X-PGP-Fingerprint: 6B 7B C3 8C 61 CD 54 AF 13 24 52 F8 6D A4 95 EF Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Tuesday, 13 June 2000 at 16:39:01 -0700, Jaye Mathisen wrote: > > I was just curious what people's thoughts were on the potential "total > system throughput" (whatever you choose it to be), wrt using software > RAID. > > It would seem that software RAID would just kill big chunks of > cache, especially CPU cache, moving lots of data throught it to > calcuate CRC's. Hmm. Yes, it might seem that way, but that's a very different order of magnitude from the I/O transfers themselves. > perhaps the raid 1/0's would be fine, since the calcuation could be > ignored. Well, there is no calculation. > It would seem that for file-server type tasks, where hopefully CPU > is not an issue, then as long as you can exceed the capacity of your > network, you're probably fine. > > And most workstations probably don't run 100% (modulo things like > the various crackers and seti), that it's a non issue. It's a non-issue anyway. The RAID-5 parity calculations don't take up any time. The real issue is that each write requires two physical reads and two physical writes, reducing write performance to about 25% of the performance of RAID-1 writes. Take a look through http://www.lemis.com/vinum.html for further details. Greg -- Finger grog@lemis.com for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message