Aight all you hardware junkies, I gots one for ya'

Machine:

Athy 1.2 T-bird, Coolermaster fan, tac reads fine, spins fine, I think its alright.

Abit K7V RAID mobo, also seems to be doing fine.

RAM: 2 sticks, 256Mb a piece, Micron PC133 SDRAM

HDA: Western Digi 7200rpm 40Gb, questionable
HDC: Maxtor 5400rpm 2.8Gb, almost definately solid
HDD: Hi-val 12x, 36x DVD-ROM, questionable
SR0: Ricoh 6x4x12 CD-RW (yeah, scsi), solid, its bloomin SCSI!

Scsi card chipset: initio i9000, take too long to probe, go figure, I've got the only non-adaptec card on my subnet.

There's obviously more gear, but its non-consequential.

Here's the issue: I bjorked (yes, that's a verb) my ext3 fs on the 40Gig'r, Slackware 8.1, and decided to re-install. The machine hung about at the end of series A.

I tried a lot of times, switching to booting scsi.s and loading from the cd-rw, opening another busybox term on F2
and running tail -f /var/log/messages. Nothing ever appeared in the tail, but occasionally I would get these weird kernel oops mentioning a virtual address, usually at 000000008 or somesuch. The drive always seemed to format right, but usually hung before it finished a bad-block checking format, always made it through the quick and dirties, and had more or less the same results under reiser, ext3, and ext2.

So what do I figure? RAM right, kernel panics in an install: hard drive, RAM, or install device? Well, I've doubled my bets by failing to install on either hard drive with either disk drive. Then I trade out the two RAM sticks and try either of them in turn alone in the board, still nada. Okay, bad install media right? Yeah, I know I used the same CD twice the day before to install Slack 8.1, but Patrick's already changed the "official" ISO and yadyadyada, and then again the CD is so free of nicks and scratches that I could shave by it.

So, try another distro... heck, I tried 4: RedHat 7.2, Mandrake 8.0, Mandrake 8.2, and Slackware 8.0. Slack 8.0 made it through to setup, as for the other three, they all errored out in the same way: except for Mandy 8.2, just once, was able to de-compress the install image to RAM (it then hung right after I selected "English" as my language), all the others hacked up while trying to do this. So, I'm back to RAM. But how in the world did I toast both sticks in one day?

I can only think of one thing: lmsensors. Six hours after I install lmsensors for the first time I got a kernel panic like the above in the middle of transfering a "very large" file, which of course leads to the downfall of my filesystem.

If anyone has a better sugestion, Pllllllllease let me know. Luckily I can perma-borrow a pair of sticks from a friend who shot his board recently (although who knows if they're any good), or yank a PC100 DIMM out of my roommate's crate.

Cheers,

Finegan

Hey Finnegan!

You're playing with all these hard disks'n'such. Make sure you have the jumper configuration correct....you know "Primary master", Primary slave etc.....

Guvvy

The only thing I popped out of the crate was the memory sticks to juggle 'em. This thing has hosted about 12 distros/OSes in the past year or so with the same drive configuration, so I'm pretty sure I nailed the jumpers.

When I get home I'm going to be trying memtest, which looks like the coolest thing since squirtable cheese, but anyone else got any ideas if my dimms turn out not to be toast?

Cheers,

Finegan

I used installed debian on a Compaq Presario...back in the day...and if I reformated the hardrive with checking for bad sectors enabled, it would never work..until I did it without checking for bad sectors and then like magic it worked, .
initio i9000?, hehe, right.
I think that your ram is fine...you didn't do anything to fry'em. Just make sure to allow many of the scsi adapters if you can't find yours exactly...I used to get a kernel panic message when I forgot to enable my adaptec scsi driver for my two ultra wide scsi hardrives. But after I recompiled the adapter in, it worked fine.

This machine panics out on SCSI (which is only connected to a cdrom), or IDE (both hard drives), and I've had some weird problems with the initio modules in the past, but plenty of those runs were made with the slackware bare.i that ignored the scsi drive entirely, so its not scsi dropping voodoo I'm pretty sure.

Cheers,

Finegan

Hi
I use memtest alot because where i work i see
lots of sick dimms pass by. One thing to remember;
memtest sometimes tells you a dimm is faulty, or
memtest itself sometimes hangs. But when one
trys another dimm it says the same. Or place the
'faulty' dimm in another board it reads ok.
I think its got something to do whith timing/caching
problems related with the mobo. So always double
check memtest results.
Another thing that leads to nasty install faults or
crashing might be the power supply.
good luck

I read the warning, one thing I'm curious about is that I haven't had issue with the memory until, well assuming its the memory, until today, over the past year and a half, so would the modules suddenly become 'incompatible'?

The PSU I hadn't even suspected... which injects a whole new hill of worry, but thanks for the pointer, I've got a furbie 300w enermax lying around to try.

Cheers,

Finegan

I was reading through the post and I was going to suggest to check the PSU - other day it drove me insane I had kernel oopses, referencing memory, al kind of weird stuff untill PSU blew up on me, such a show with smoke and smell, luckly my other hardware survived the attack! After I replaced PSU no more oopses, memory references and blah.

maybe the the word 'incompatible' fell on the wrong thing
(yes thats bad english i know )

i meant memtest, sometimes being incompatible with
certain mobo's, not your mobo being incomp. with
its own home placed mem stickies

Thanks for the clarify Mr.Gee, I checked and I've got one of the more vanilla via chipsets ever produced. It made the first run without a single memory error, and even tested my friend's RAM as 1 stick good, 1 stick bad, so I think it did its job right, killer friggin' utility there! just awesome.

I switched PSUs, same results.

I disconected the IDE devices here and there, same results...

I'm now trying something that hadn't even occured to me at first... After installing lmsensors the other day, I tried out SANE, and in so doing changed BIOS parport mode. Now thinking somehow on the planet that might have been it, I changed it back last night early on, but then I tackled something I hadn't even thought of...

Reset BIOS to default values.

Why not, eh? When changing things before to get ecp parport mode to work, I must have changed something else, becuase I'm mid-way through installing package K and the machine is just flying along.

Thanks again for all of the help that poured out in the span of a few hours. I knew there was a reason I liked this place.

Cheers,

Finegan

The old buggy bios strikes again....I get all sort of errors if I try any bios newer than ZT in my KT7A....go figure....

Good to hear it works again
and finding the answer yourself is always
great for characterbuilding

had my own troubles with bios when installing freebsd
4.5/4.6 onetime. It booted alright, didn't remount the
cdrom in second stage; had to flash my bios and things
went as they should.