Odd Problems

[Dun'kalis]

I've been having a TON of problems with my hardware lately. My system:

Athlon 1.4GHz
768MB RAM
nVidia GeForce 2 MX400
Creative Modem Blaster ISA
RTL1839
Maestro3 Sound Card

IDE Channel One
80GB 5400RPM HDD
5GB HDD (disconnected)

IDE Channel Two
CD-RW
DVD-ROM

The problem is near-constant lockups. Its most likely hardware, but I'm not too sure. It happens more in Linux, but it happens in Windows. Thats why it is most likely hardware, and Windows is more fault-tolerant.

I'm going to try memtest86 (may have some bad memory modules), and I'm hoping I don't have bad memory...Anyone else have any ideas?

EDIT: memtest86 came up with...no errors. I then proceeded to boot to Windows (since its more fault-tolerant than Linux), and it took about 10 attempts to boot. To get some more troubleshooting information, I checked my Linux kernel logs, and I found this tidbit I had never seen before:

Spurious 8259A interrupt: IRQ 7

And, on one boot attempt, POST gave me this fun error:

CMOS Checksum bad. Resetting to defaults.

I almost wish it was the memory...

[pauper]

First off, check your CMOS battery voltage (should be at 3.0 or better, if not, change it)

Clean out the inside of the computer case, get rid of the dust bunnies etc. (if you haven't already)

While you're in there, have a look at the parallel port connections, make sure they aren't bridged. If you have a printer connected, disconnect the cable from the PC. (I'm working off the irq7 interrupt clue on this one)

The 8259A is one of two interrupt controllers that are bridged together. Since the bridge is irq1 to irq9, then a spurious interrupt at irq7 can only be from something connected to that irq (normally lpt1) or a fault within the chip itself. Hopefully the above will eliminate anything external to the chip and even more hopefully the problem itself!!

Fingers crossed! :)

[wygargoyle]

I can't say that your lockup problem definitely is not related to the "spurious interrupt" message, but those messages seem to be fairly common with epro100 and 8139too NICs. I use those 2 types of cards almost exclusively and see at least one of those messages just about every time an address is assigned to the card (i.e. during network startup). I would look elsewhere for your lockup problem.

[Dun'kalis]

In Windows, I checked my IRQs...There is nothing on IRQ 7. I've got nothing installed in my parallel port. I forgot how to check the IRQs in Linux (I looked everywhere in /proc/...)

Voltage is fine...

I grabbed a flashlight, and found a spot covered with dust...I'll clean it off in a bit. Its around the processor socket.

One thing I'm thinking of is either a flaky or inadequate power supply. Its 300W, which may not be enough. The person who built the computer says its enough, but, then again, I don't exactly trust him with computers (sadly, other people do...)

EDIT: I was looking through Windows logs, and at the time of virtually every crash, there are 4 ACPI error messages...Hmm...Trouble is, when I disable ACPI, Windows refuses to boot.

[wygargoyle]

Your interrupts are listed in /proc/interrupts and you can get more info about what is attached to them by looking in /proc/pci. I also agree that it's quite possible that your power supply is not quite enough to handle what you have in that box, but seems that more often weak power supplys will cause reboots rather than lockups. At any rate, I found this powersupply guide the other day and it has some very useful information that might help you figure out if your powersupply is hefty enough for the job.

http://firingsquad.gamers.com/guides/power_supply/

[Qwijibrumm]

What motherboard do you have? You have a 1.4Ghz T-bird which is noteably unstable when coupled with a NVidia card. If you have a KT-266 based motherboard on top of it that would really explain things. The KT-266 chipset was unstable as hell when it was released (again especially when coupled with the 1.4GHZ T-bird). If you have an old BIOS revision then a BIOS update might be in your future.
And if your Motherboard happens to be an MSI K7T266 just buy a new processor, because you will never get that combination stable.
Don't get me wrong, none of these components are bad. They are all high quality. But, there was a freak incident and none of these work well together for some reason.
Oh and your power supply should be sufficient, I was running a similar combination with no problems. (But my friend with the components I listed....)

[Dun'kalis]

My motherboard is a Tyan Trinity KT-A (VIA KT133-A Chipset).

Anyway, I shut it down, blew out some dust, and, well...It wouldn't boot.

Then, I spent a few hours doing other stuff, came back, turned on the PSU (separate switch for the PSU, but usually also boots the machine). Didn't work, so...I hit the switch on the front, and it booted. Go figure.

Anyway, I'm gonna reboot to Linux after I'm done here and see if the problem still exists...

The PC is only 2 1/2 years old...