From owner-freebsd-questions Tue Apr 14 13:59:25 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id NAA11801 for freebsd-questions-outgoing; Tue, 14 Apr 1998 13:59:25 -0700 (PDT) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from top.worldcontrol.com (surf52.cruzers.com [205.215.232.52]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id UAA11362 for ; Tue, 14 Apr 1998 20:58:19 GMT (envelope-from brian@litzinger.com) From: brian@litzinger.com Received: (qmail 12532 invoked by uid 100); 14 Apr 1998 20:59:20 -0000 Message-ID: <19980414135915.A12460@top.worldcontrol.com> Date: Tue, 14 Apr 1998 13:59:15 -0700 To: bh@epigram.com, freebsd-questions@FreeBSD.ORG Subject: Re: RAID performance/benchmarking Mail-Followup-To: bh@epigram.com, freebsd-questions@FreeBSD.ORG References: <3533BDAF.A8588F53@epigram.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.91i In-Reply-To: <3533BDAF.A8588F53@epigram.com>; from Brandon Huey on Tue, Apr 14, 1998 at 12:49:03PM -0700 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On %M 0, Brandon Huey wrote: > Dell PowerEdge 4200/300, 96MB RAM, 3 9GB Seagate Barracuda SCA drives > DPT PM3334UW w/64MB cache > > FreeBSD 2.2.6 > DPT kernel options: DPTOPT > DPT_MEASURE_PERFORMANCE > DPT_FREELIST_IS_STACK > > using IOZONE and Bonnie, i create a 128MB file for read/write > benchmarking. > > i am aware of the performance implications of RAID-5, but I am getting > about 2MByte/sec max write transfer rate compared with 19MByte/sec on > reads. In RAID 5 to write a sector requires the following: (through not necessarily in the order specified) read the parity sector (usually on a different drive than the target sector) read the target sector write the target sector write the parity sector while a read only requires a read of the target sector. So generally writing takes quite a hit. -- Brian Litzinger To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message