From owner-freebsd-hackers Thu Aug 31 05:57:50 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.FreeBSD.org (8.6.11/8.6.6) id FAA04584 for hackers-outgoing; Thu, 31 Aug 1995 05:57:50 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.FreeBSD.org (8.6.11/8.6.6) with ESMTP id FAA04578 for ; Thu, 31 Aug 1995 05:57:48 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id FAA11224; Thu, 31 Aug 1995 05:57:14 -0700 From: "Rodney W. Grimes" Message-Id: <199508311257.FAA11224@gndrsh.aac.dev.com> Subject: Re: 4GB Drives To: pete@kesa26.Kesa.COM (Pete Delaney) Date: Thu, 31 Aug 1995 05:57:14 -0700 (PDT) Cc: pete@kesa26.Kesa.COM, jbryant@argus.iadfw.net, freebsd-hackers@freebsd.org, pete@rahul.net In-Reply-To: <9508311100.AA05592@kesa26.Kesa.COM> from "Pete Delaney" at Aug 31, 95 04:00:52 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 3270 Sender: hackers-owner@freebsd.org Precedence: bulk > > > So Rodney, which 4 GB Drive would you buy? Today, if you ordered a 4G drive from me I would first try to talk you into 2 2.2G drives to load balance if this was a single disk for a single machine: XX. TMG QTM11265 Quantum, Capella 2.2GB, 3.5"x1", 5400RPM, 8.5MS $ 800.00 XX. TMG QTM11265 Quantum, Capella 2.2GB, 3.5"x1", 5400RPM, 8.5MS $ 800.00 If you already had disk space, or where looking for multiple 4G drives to load balance your news spool on I would tell you: XX. TMG SG15230N Seagate 4.3G, 3.5"xHH, SCSI-II, 5400RPM, 9mS $1232.00 Yes, that is correct, Accurate Automation is actually _RECOMMENDING_ a SeaCRATE drive. But only this ONE model at a single point on the size curve. If you must have speed (and you are going to bleed to get it) I would recommend: XX. BAS HDMC3243W Micropolis MC3243W 4.23G 3.5"xHH 7200RPM 8.5mS $1495.00 [Also avaliable in non-wide, just not on my quote clip cheat sheet :-)] This is a little public presentation of what seperates me from the others in this market. When systems are requested for quotation from me I see things like ``4G disk''. Well, that one item becomes a whole pile of questions about just what you need 4G for, and what performance levels you are looking for. As you can see I can tune that 4G to fit a wide range of price/performance points. I am not the clone shop that has 1 solution for all problems, I have N configurations to fit the M problems, all machines are custom quoted based upon rather length email iterations in most cases. Just ask any of my customers, it kinda shocks a few of them when there request for quotation comes back with a 2 page email asking all sorts of details about what it is they are doing with that 8G of disk, and ``Penitum'' with an ethernet and SVGA. In the news server class of machines I have 3 proposals before me right now, they all look just a little different in disk configurations due to just how these new servers are going to be used to distribute news. Some are what I would call and endpoint news server, like I would install here at the end of a T-1 to provide news service for a reasonbale size company, others are central servers that will have T-1 pipes to the internet and downline feed news to other sites via T-1 and below bandwidths, these go into your typical mom and pop ISP. The third class is the news ultra server class clusters with above T-1 bandwidth (upto DS-3), that distribute news to whole regions, these have some rather high performance disk system reqirements as well as some expensive network gear and often require multiple machines to provide the horse power (thus the term ``cluster''). Now, can you all leave me alone for 30 days so I can go get the stripes working, I have small bottleneck that needs fixed :-):-):-) And can anyone tell me what the mean and standard deviation of an I/O request to an aic7870 is before it hits the drive given 0 scsi bus contention? This seems to greatly effect rotation offset on stripe sets when pushed to the limits of data coming under the head just after the I/O hits the drive. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD