From owner-freebsd-stable Fri Oct 13 0: 5: 0 2000 Delivered-To: freebsd-stable@freebsd.org Received: from peak.mountin.net (peak.mountin.net [207.227.119.2]) by hub.freebsd.org (Postfix) with ESMTP id 495AA37B502 for ; Fri, 13 Oct 2000 00:04:56 -0700 (PDT) Received: (from daemon@localhost) by peak.mountin.net (8.9.1/8.9.1) id CAA25267; Fri, 13 Oct 2000 02:04:53 -0500 (CDT) (envelope-from jeff-ml@mountin.net) Received: from dial-106.max1.wa.cyberlynk.net(207.227.118.106) by peak.mountin.net via smap (V1.3) id sma025264; Fri Oct 13 02:04:39 2000 Message-Id: <4.3.2.20001013005356.00ba4a20@207.227.119.2> X-Sender: jeff-ml@207.227.119.2 X-Mailer: QUALCOMM Windows Eudora Version 4.3 Date: Fri, 13 Oct 2000 02:04:46 -0500 To: Carroll Kong From: "Jeffrey J. Mountin" Subject: Re: Bad IDE Drive Cc: Leonard Chung , freebsd-stable@FreeBSD.ORG In-Reply-To: <4.2.2.20001012192324.00c0de90@email.eden.rutgers.edu> References: <4.3.2.20001010190749.00c5cf00@207.227.119.2> <4.3.2.7.2.20001010124858.026637c0@yikes.com> <200010100509.XAA18135@harmony.village.org> <4.3.2.7.2.20001009190324.028c6d58@yikes.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG At 07:58 PM 10/12/00 -0500, Carroll Kong wrote: CC's trimmed. >>This smells like a techie sales pitch. Little meaningful info regardless >>of terms and large words. Did I mention I'm a hard sell? > >Actually, it is probably true. On the low-end, IDE's although rather rare >bursts, would most likely outperform an older scsi drive. I got a 9.1 G >Ultrastar 9LP and a maxtor 27 gigabyte ide drive. Sorry, the maxtor >usually wins out. My point was the mixing of various terms, cost, and no clear application. For some applications the latest IDE "king" can beat out SCSI's "king" (incidentally the Atlas2 regained the crown from the FXP75). >I do not believe there is a strong correlation as to "ide drives dying >faster" than "scsi drives." In fact, from my personal experience, I >believe I have felt the same if not more leaning towards the IDE >drives. I still vote for SCSI for the most part though. This may change, but if indeed the perception is wrong, then only time will tell. Some SCSI drives have spent the better part of a decade being thrashed from various people's testimony. Personally I had 3 lightly used IDE drives die within a 2 week period or was it 4 in a 3 week. Different brands and models as well. IMO, WD went to pot years back their choice on concentrating on the low end market doesn't help. Seagate's IDE drives seem to have been among the slowest around. Years back started having better results with Maxtor and then IBM came in with the best performance and price for *both* SCSI and IDE. Many have talked about their good experiences with the former, but can't say I recall much on the latter. With SCSI never liked Fujitsu much (unless I wanted to fry eggs). Seagate was good, but Micropolis started doing better. Never even thought about WD's lame attempt. IBM's been the choice for a while. Everyone's opinions vary, but what seems clear is that IDE has gained far more than SCSI has performance wise. However, in most cases they do seem to fail more often, but not as much as they used to, IMO. >Well, not sure how old you are comparing your "slower" scsi drives... just >telling you from personal experience .... oh yeah you want numbers? Ah yes, my "newer" IDE boot drive does fairly well compared to the "older" SCSI drives. Until I start doing many things at the same time. Between new and old is a few years and to be honest, the newest is at least 3 years old. >www.storagereview.com Numbers for single drive testing mean little in a server context. Sure the newer IDE can lay down large files much faster, but doing any disk IO during a buildworlds and the SCSI wins. Drives aren't the issue, but architecture is for several reason. >Ignore the crap with the other benchmarks. The IoMeter benchmark (by >intel) is the better test. If you got an older cheetah or barracuda, you >will probably get comparable if not superior performance against the new >ides. If you got anything less... it's almost definitely slower. Of >course the drivers will make a difference for cpu utilization. (scsi >usually wins in this one). My point to Leonard off-list was to use a FBSD setup for testing. Nothing large scale. Single drive and a 4 drive RAID 0+1 setup would be sufficient. More pertinent to the list and less likely to be met with suspicion coming from M$. >Finally, I am neither an IDE fan nor SCSI fan in all situations. They >have their places in different areas. For Dollar per Megabytes, you >really win with IDE. For performance, you might even say so >(economically). For the absolute best... SCSI wins out. As for the raid >situation, you'd need a dedicated channel for each IDE drive... good lord, >does that thing take up IRQs per channel? If so, cannot possibly scale to >match up to a good scsi raid. Usually the "boot" drive is IDE and SCSI handles everything else. Have had good luck with this with zero failures (knock on wood). That is partly due to rotating out drives from servers to workstations when doing drive-swap upgrades, but I do have a couple that have been around for a while. Been wondering for 2 years when one is going to die. By it's sound, anytime. Not that I want it to. Call it perverse, but am facing an hour drive at any time. Murphy's Law - It's expected to fail and therefor will not. Recently the CPU fan crapped out instead. Now to get a little more on topic here... Most likely going to checkout the 3Ware controller. Anyone running these on -stable? From Mike Smith's pages, some have deployed them in production environments and that was in late May. A more appealing IDE solution than boards with 4 channels and needing 4 IRQ's. Only issue is case layout, as mentioned elsewhere. If it pans out, will be able to save one party about half the cost of hard drives in a new server. And if not... I wanted to pickup some new hardware. ;) Gimme a year or three and it might be considered for low end solutions. Otherwise, it's an alternative with no guarantee. Jeff Mountin - jeff@mountin.net Systems/Network Administrator FreeBSD - the power to serve To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message