From owner-freebsd-hardware Wed Aug 22 11: 0:59 2001 Delivered-To: freebsd-hardware@freebsd.org Received: from krell.webweaver.net (krell.webweaver.net [64.124.90.11]) by hub.freebsd.org (Postfix) with ESMTP id ECD4337B40F for ; Wed, 22 Aug 2001 11:00:01 -0700 (PDT) (envelope-from nicole@unixgirl.com) Received: from xwin.nmhtech.com (xwin.daemontech.net [208.135.51.161]) by krell.webweaver.net (Postfix) with ESMTP id 19B6D20F2E; Wed, 22 Aug 2001 11:00:00 -0700 (PDT) Message-ID: X-Mailer: XFMail 1.4.0 on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <200108221652.JAA03752@mina.soco.agilent.com> Date: Wed, 22 Aug 2001 11:00:00 -0700 (PDT) From: Hodge Podge To: Darryl Okahata Subject: Re: 3ware stuff not ready for heavy duty useage-followup Cc: Michael VanLoon , freebsd-hardware@FreeBSD.ORG, Marc Rassbach Sender: owner-freebsd-hardware@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org On 22-Aug-01 Darryl Okahata wrote: > Hodge Podge wrote: > >> Yes however I will bet even money that if I put a seperate power supply on >> every friggen drive and still have the same problem someones gonna say.. I >> bet >> one of Those supplys was bad.. Can folks try, just try to imagine that they >> "could" be wrong and come up with some other reason? Do I honestly have to >> Prove it wrong in some scientific way before you are willing to imagine some >> other possibility? Jeeze. Some time I or someone else will have a chance to >> test the theory, but for now that is all it is. One thoery. > > OK, let's look at the other possibilities: > > 1. Non-3ware hardware problem (bad motherboard, RAM, etc.). > > Very unlikely, as this problem happens on "multiple systems". Ny belife as well. > 2. Bad batch of 3ware boards. > > Unlikely, but a very small sliver of a possibility. Also true. > 3. 3ware hardware bug. > > If so, you're screwed. There's probably nothing you can do about it. Very True > 4. 3ware firmware problem. > > If so, you're probably screwed. At the very least, you'll probably > have to wait some unknown, probably long, period of time for a > firmware update. My points. Everytime I loose a drive to the card, The drive tests fine elseware for other usage such as in JBOD. Also the continued failure of the system to be able to rebuild onto the hotspare is very bad. Also loosing 2 drives on the same plex (under RAID 10) preventing any hope of a rebuild points to the drives in many ways and how the card is interacting with them. > 5. FreeBSD problem. > > Not very likely, but a possibility. Other people seem to be having > success with 3ware and high loads. I have not seen anyone mention > using an 8-drive setup, though (I get to do this in a few weeks, > although the system won't generally be under high load). > > Unless you're conversant with FreeBSD disk driver debugging, you're > at the mercy of the FreeBSD driver developer (yes, I know who it is), > whom you may have unintentionally pissed off beyond redemption. At > the very least, in order to fix any problems, he'd have to reproduce > your problems. If I have somehow pissed off Mike I appologise. But that seems like pissing off the weatherman for complaining abt the rain. I can't see how this could have anything to do with the driver since the interaction causing the disks to dropout is between the card and the drive. Mike, if I have somehow managed to piss you off with this, please let me know how so I can correct that. The whole reason for my posting in the first place was.. Hey I have been having problems like mad, here is my reasoning why... be carefull and test a Lot before you rely on a similiar system. I don't think the Larger drives are ready for prime time under heavy loads. If it works for you.. great.. But be careful. > [ Note: as another person has mentioned, you're coming across as a > real hard case (even before your recent messages). I'm sure it's > unintentional, but you've got to be aware of this. If you want > help, you don't want to (unintentionally) piss off the people who > might be able to help you. Please note that many of the people > here are anal-retentive analytics who need data, data, and more > DATA (yes, I'll raise my hand ;-). You haven't given much details, > and, for much of the details which you have given, you've given as > part of "snippy"/"unintentionally insulting" responses or > maddeningly dribbled them out across many postings. ] Well, At first I was amazingly pissed off for loosing yet another weekend to having to rescue a system. Some of that may have come off incorrectly. I appologise if that has occured. I also do not have time to write a dissertation on the issue. Perhaps the details below will help. System1 - iteration 2 4U case with 300 Watt power supply AMD 900 Mhz proc on an ASUS MB (don't remember the model) 4 40 Gig Maxtor hard drives in RAID 10 3 40 Gig Maxtor hard drives as hot spares S3 video card - teac floppy - Intel dual port ether card 1 6800 series 8 port 3ware card with latest firmware IDE cables that came with the 3ware card. System 2 - iteration 2 4U case with 400 Watt power supply AMD 1.1 Ghz CPU Iwill 266 MB S3 video card - intel dual ethernet card - Netgear GB copper card 8 Maxtor 61 GB 7200 RPM hard drives - 7 in RAID 5 1 Hot spare 1 7800 8 port 3ware card with latest firmware IDE cables that came with the 3ware card. Iteration 2 means the 2nd attempt setup with the 3ware card. Prior itorations had the IBM IDE drives and slightly different configurations. > 6. Large noise sources near your systems. > > Unlikely, as you haven't mentioned any special situations (like > having an arc welder, large motor, or transmitter near your systems). Unlikly in the datacenter. > 7. Bad line power. > > Unlikely (well, unknown, as you haven't given any details), but, as > Borja mentioned, a bad UPS can really screw you over. Hmm Well Supposedly the power In AboveNet is pretty good. > 8. Power supply problem. > > Unknown. Seems pretty likely, but you're also seeing this problem > across multiple systems, which can imply that it's not a power supply > problem -- however, you haven't said that any of the other systems > have large, beefy power supplies. > > This is one of the few things that you can test and control. Somewhat yes. My time however is also being directed into building a SCSI system. The only problem we have ever had with any of the SCSI systems is when one of the brain dead zombies at abovenet hit reset on the wrong server for us. (I could tell you some horror stories from having to use ABN "remote hands" assistance) So far I would have had better luck with Mr Magoo. > Well, the above are the possibilities as I see them. It's your choice. > What do you want to do? Well since someone else has build a system that can be used for testing, I would be interested in what you and the others would like to do as a test. I do not belive he would have a sampling osciliscope of illrepute available to test for minute spikes so what would make people happy there? As far as what can be used to pound on the system that he has available I am uncertain. Certainly running the system as I had in RAID 5 with one hot spare would cause all the disks to work at once and figuring some way to feed it lots of data to write to the disks would be the best. Perhaps ideally, it would be nice if there would be a way to see what error messages are being sent to the 3ware card and how the card is dealing with them. What actions might create what errors and what occurs durring the rebuild process. Hope this is more of what you wanted. Nicole > -- > Darryl Okahata > darrylo@soco.agilent.com > > DISCLAIMER: this message is the author's personal opinion and does not > constitute the support, opinion, or policy of Agilent Technologies, or > of the little green men that have been following him all day. > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-hardware" in the body of the message ******* |\ __ /| (`\ ******* * * | o_o |__ ) ) * * * * // \\ * * * Nicole Harrington | AKA Hodge Podge * ----------------------(((---(((-------------------------------- nicole@unixgirl.com http://www.unixgirl.com/ webmistress@dangermouse.org http://www.dangermouse.org/ nicole@deviantimages.com http://www.deviantimages.com/ -- Powered By Coca-Cola and FreeBSD -- "The world is run by those who show up" -- The Best Place for Your Web Site - www.WebWeaver.net -- ---------------------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message