Date: Sat, 22 Apr 2000 22:14:52 -0700 From: Alfred Perlstein <bright@wintelcom.net> To: Martin Nilsson <martin@filex.se> Cc: freebsd-stable@freebsd.org Subject: Re: amr still seems to have issues. Message-ID: <20000422221452.D204@fw.wintelcom.net> In-Reply-To: <390236EC.80A83E0B@filex.se>; from martin@filex.se on Sat, Apr 22, 2000 at 11:34:04PM %2B0000 References: <390236EC.80A83E0B@filex.se>
next in thread | previous in thread | raw e-mail | index | archive | help
* Martin Nilsson <martin@filex.se> [000422 17:04] wrote: > * Alfred Perlstein <bright@wintelcom.net> [000421 10:57] wrote: > > > > > > Apr 15 13:31:06 abacus /kernel: amr0: command 31 wedged after 3= 0 seconds > > >=20 > > > Hmm. Looks like I'm not retiring the wedged command correctly. This= is=20 > > > a symptom, rather than the real problem, though. See if disabling th= e=20 > > > command timeout stuff makes the system happy - either these commands = are=20 > > > just taking a _long_ time to complete, or you have another problem. > >=20 > > Yes, I've been thinking I may need some firmware upgrade from AMI? > >=20 > > I'll look into it. > >=20 > > Ok, the latest firmware is the uf80-1.61, which is on the card. >=20 > The latest firmware for the series 428 cards is UF82. Get it from > www.ami.com grrrr... the tables were so fubar'd under netscape that I didn't see it. I'll be upgrading it asap. thanks! > > > I _still_ think you have disk, cable or enclosure issues, but now tha= t=20 > > > you've precipitated this case I can go look at what I'm doing wrong h= ere.=20 > > > Thanks! >=20 > =03> I was wondering if anyone else on the lists have had the same > problems > > as I have been having with the Enterprise 1200 (3 scsi port) version. >=20 > I have a series 428 (Enterprise 1200) with two channels and 32MB RAM in > a=20 > Tyan dual PPro 200MHz running 5.0-current. This is not a production > machine, > and the are no important data on the disks so I can experiment with it > as I like. > =20 > I've had all kinds of problems with it and the three Quantum Viking > disks I have. > It trurned out that i must disable SCSI command Queing on the disks in > the AMI BIOS > otherwise the controller stops working within a couple of minutes. I don't have this problem... well it takes a heck of a lot longer than a couple of minutes... 9:43PM up 2 days, 5:11, 2 users, load averages: 0.01, 0.03, 0.00 > This error is more reproducible using RAID0 than RAID1 or 5.=20 > I've also had big trouble with a SCA to Wide and narrow converter on one > of the=20 > disks. My experience is that the NCR SCSI controllers (as on the AMI > cards) are=20 > rather picky about bad SCSI cabling and termination, esp. when running > Ulra Wide. Yes, this concerns me, I may yank the box to check the job they did with the scsi cabling... > > I'm really not using all that much disk throughput. > > I'll have that iostat log the next time it keels over. >=20 > I can make it fail regerdless of thruput, this is not the old controller > wedged bug. >=20 > I have just put i thru two make worlds (-j64) without any problems, two > disks RAID0 > one on each channel. First time with default parameters, second time > with Write back, > adaptive read, cachedIO.=20 Wait a sec, you _can_ or _can not_ get it to crash with the latest firmware and turning off command queuing? > Tell me if you want me to do something special with it to test it. At > present I can=20 > only run it with two disks as I only have two reliable hotswap cradles, > I have=20 > access to three more disks but I must buy more cradles before I can use > them.=20 I really can't think of anything except getting us a traceback if you can make it panic. thanks a ton! --=20 -Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org] "I have the heart of a child; I keep it in a jar on my desk." To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000422221452.D204>