From owner-freebsd-hackers@FreeBSD.ORG Wed May 21 21:06:46 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 879CE37B401 for ; Wed, 21 May 2003 21:06:46 -0700 (PDT) Received: from backmaster.cdsnet.net (backmaster.cdsnet.net [63.163.68.2]) by mx1.FreeBSD.org (Postfix) with SMTP id B5B9D43F85 for ; Wed, 21 May 2003 21:06:45 -0700 (PDT) (envelope-from mrcpu@backmaster.cdsnet.net) Received: (qmail 83662 invoked by uid 29999); 22 May 2003 04:06:43 -0000 Date: Wed, 21 May 2003 21:06:43 -0700 From: Jaye Mathisen To: hackers@freebsd.org Message-ID: <20030522040643.GS75453@backmaster.cdsnet.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i Subject: Abysmal mly Extreme RAID performance in 5.0-p7? X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 22 May 2003 04:06:46 -0000 Have a p2-450 with a mylex 3000 fibre channel RAID card. Configured a few different kind of arrays on it, however, the one I'm interested in right now is this one: da1 at mly0 bus 3 target 1 lun 0 da1: Fixed Direct Access SCSI-3 device da1: 135.168MB/s transfers da1: 274656MB (562495488 512 byte sectors: 255H 63S/T 35013C) It's 16 10kRPM seagates striped in a RAID0. This is the best # I've gotten so far with this controller combo. 206438400 bytes transferred in 27.405360 secs (7532775 bytes/sec) mly0: port 0xc800-0xc87f mem 0xd8000000-0xdfffffff,0xd4000000-0xd5ffffff irq 12 at device 8.0 on pci2 mly0: controller initialisation started mly0: eXtremeRAID 3000, 3 channels, firmware 7.00-3-00 (20011219), 128MB RAM da0 at mly0 bus 3 target 0 lun 0 da1 at mly0 bus 3 target 1 lun 0 da2 at mly0 bus 3 target 2 lun 0 Interestingly enougn, I haven't tried any plain SCSI disks on a 5.0 box, but I notice the lack of a "tagged" scsi device entry anywhere for those disks... So I"m thinking the no TAGGED QUEUEING or whatever it used to be under 4.x might be related to the issue... The arrays were all initialized in the BIOS, and the drives were idle before starting the test. No background stuff going on. ANy tips appreciated.