From owner-freebsd-stable Mon May 27 6: 3: 3 2002 Delivered-To: freebsd-stable@freebsd.org Received: from point.osg.gov.bc.ca (point.osg.gov.bc.ca [142.32.102.44]) by hub.freebsd.org (Postfix) with ESMTP id DB60B37B403 for ; Mon, 27 May 2002 06:02:57 -0700 (PDT) Received: (from daemon@localhost) by point.osg.gov.bc.ca (8.8.7/8.8.8) id GAA02388; Mon, 27 May 2002 06:02:44 -0700 Received: from passer.osg.gov.bc.ca(142.32.110.29) via SMTP by point.osg.gov.bc.ca, id smtpda02386; Mon May 27 06:02:33 2002 Received: from cwsys.cwsent.com (cwsys2 [10.1.2.1]) by passer.osg.gov.bc.ca (8.12.3/8.12.3) with ESMTP id g4RD2OLS005773; Mon, 27 May 2002 06:02:24 -0700 (PDT) (envelope-from cy@cwsent.com) Received: from cwsys (localhost [127.0.0.1]) by cwsys.cwsent.com (8.12.3/8.12.3) with ESMTP id g4RD2Nr1099239; Mon, 27 May 2002 06:02:23 -0700 (PDT) (envelope-from cy@cwsys.cwsent.com) Message-Id: <200205271302.g4RD2Nr1099239@cwsys.cwsent.com> X-Mailer: exmh version 2.5 07/13/2001 with nmh-1.0.4 Reply-To: Cy Schubert - CITS Open Systems Group From: Cy Schubert - CITS Open Systems Group X-os: FreeBSD X-Sender: cy@cwsent.com To: Kirk Strauser Cc: freebsd-stable@FreeBSD.ORG Subject: Re: Hardware RAID vs vinum In-Reply-To: Message from Kirk Strauser of "26 May 2002 22:52:33 CDT." <87r8jy9ri6.fsf@pooh.int> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 27 May 2002 06:02:23 -0700 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG In message <87r8jy9ri6.fsf@pooh.int>, Kirk Strauser writes: > > At 2002-05-27T03:41:39Z, "Phil Rosenthal" writes: > > > Is there any reason why a hardware raid would be better, am I missing out > > on anything? > > In theory, a hardware RAID should give you better CPU utilization. After > all, a seperate processor (or DSP, etc.) would be doing all of the work that > your main CPU is currently doing. Furthermore, your CPU would be able to > forget about all of the extra abstractions of Vinum, such as plexes, > subdisks, and so on - the RAID could look like one single harddrive. > > In reality, I don't think that the RAID controllers would protect the CPU > from nearly so much work, and I'm almost positive that your CPU will be > orders of magnitude faster than the processor on the controllers. I beg to differ. I don't have any experience with vinum, however I do have experience with other products such as VxVM (Veritas Volume Manager) and ODS (Online Disk Suite). About three years ago a rumor was floating around the building that "NT was faster than Solaris." The customer in this case had a Sparc Ultra 2 system using RAID 5 on ODS (software RAID 5). The application people and the NT folks managed to prove that when the Oracle application was run on an NT system using hardware RAID 5 (of course they didn't mention the hardware RAID 5 part), it ran 37 times faster than on the Sun system. Sun found out and quickly shipped an Ultra 2 with an A3500 array to prove that Sun was no slouch. It turned out that moving the application to hardware RAID improved performance by 40 times. (In the end the customer was not willing to spend the money on an A3500 array so to this day the application still uses software RAID 5, but we've not installed software RAID since then.) Did we see a difference in CPU utilization? No. However ODS and VxVM run below the kernel and cannot be measured by vmstat or sar, which is different to how vinum works because it is part of the kernel. The moral of the story: You get better mileage out of a hardware RAID controller. -- Cheers, Phone: 250-387-8437 Cy Schubert Fax: 250-387-5766 Team Leader, Sun/Alpha Team Email: Cy.Schubert@osg.gov.bc.ca Open Systems Group, CITS Ministry of Management Services Province of BC FreeBSD UNIX: cy@FreeBSD.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message