From owner-freebsd-current Thu Oct 30 17:35:51 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA28553 for current-outgoing; Thu, 30 Oct 1997 17:35:51 -0800 (PST) (envelope-from owner-freebsd-current) Received: from sendero-ppp.i-connect.net (sendero-ppp.i-Connect.Net [206.190.143.100]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id RAA28537 for ; Thu, 30 Oct 1997 17:35:44 -0800 (PST) (envelope-from shimon@sendero-ppp.i-connect.net) Received: (qmail 23726 invoked by uid 1000); 31 Oct 1997 01:36:05 -0000 Message-ID: X-Mailer: XFMail 1.2-beta-100797 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit MIME-Version: 1.0 In-Reply-To: <34591A3C.B81619A2@excite.com> Date: Thu, 30 Oct 1997 17:36:05 -0800 (PST) Organization: Atlas Telecom From: Simon Shapiro To: kingson@excite.com Subject: Re: Any luck with DPT SmartRaid 4 and SMP FreeBSD-current??? Cc: freebsd-current@FreeBSD.ORG, Tom Sender: owner-freebsd-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Hi Kingson Gunawan; On 30-Oct-97 you wrote: > Hi Tom, > I am planning to use RAID 0 (no mirroring), just a straight stripping. > BTW, potentially, I will need to grow the array up to around 120GB. > However, that's somewhat in the future. > This array will be read intensive. Read performance improves with array size. Write performance will degrade with array size (Raid-5), but the rate of degradation is not linear, nor exponential. Expect to write about 5-6MB/Sec. Read at double or better. We max out around 1740 RAID I/O operations per second. Compare that to other controllers. Also keep in mind that ccd consumes processor memory and cycles, while a DPT will do it all on-board with exactly one interrupt per access. A CCD striped four wide will generate four times as many interrupts. Luck? We are shipping FT transaction processing systems based on exactly that. Uptime is in the weeks (we stop for upgrades :-), and no known crashes. Zero data loss. An IMPORTANT (again, VERY IMPORTANT) factor in successfull SCSI array is quality of disks, enclosures, cabling and power. BTW, the DPT will do ECC over its own memory, all the way into the disk sector, so bus glitches, minor errors, etc. are corrected in real time. If you have 120GB on-line, that may be important. We have only 16GB on entry level systems but the monetary value of the database is well worth the expense :-) --- If Microsoft Built Cars: Every time they repainted the lines on the road, you'd have to buy a new car. Sincerely Yours, Simon Shapiro Atlas Telecom Senior Architect 14355 SW Allen Blvd., Suite 130 Beaverton OR 97005 Shimon@i-Connect.Net Voice: 503.799.2313