Linux isn't supposed to crash they say, eh?
Well anyway I would like to get some help on this problem I have with my RHL9 box. It has come to my attention that sometimes my RHL9 box stops responding out of the sudden. At first this happened only while I was away from the PC itself and you must know that I'm almost always away because I use my linux PC as a router + firewall for my internal network so if this PC dies, noone can access the internet. I only do occasional checks to see that everything is running smoothly.
When it dies I can usually tell by the internet not working and that my RHL9 screen remains black (or the last picture that was visible before the crash if the screen saver doesn't pop up or the monitor turns itself off) when I try to move the mouse.
This gets irritating if my PC freezes in the mid-week when I'm not at home so my family has to wait for me to come back and reset the things so that they can access the internet once more. But recently it has come to my attention that PC will freeze mostly when there is high CPU load on it. I tested and confirmed this by loading a web browser and playing a Flash game in high quality which caused a high CPU load that I can tell since my internal CPU fan rises its tone for it spins faster when there is increased CPU load. It crashed 3 times in a row during the game! Hattrick!
Now I would like to know what causes it to crash. Hardware or software fault? What to do to fix? At first I thought it was my bad RAM but I tested it with GoldMemory and it gave a green light for the RAM.

PC specs (yes, it's an old PC):
VIA Tech. motherboard (VA6)
Pentium 2 300MHz
256 MB SDR RAM
3Dfx Voodoo 3 2000
40GB HDD (seagate)
250W PSU
CD Rom
Floppy drive

Thanks in advice.

EDIT: Pc specs updated

I forgot to mention that I also noticed occasional glitches while running the flash game. They seemed to follow periodically and made my PC freeze for a moment and then continue working as normal. This behaviour continued all the way until the "final" glitch that froze up the PC for good.
If this is a software bug then I might want to change to a different distro since I wish to try some others out. The bad news would be a hardware error. I didn't have this bug all since the beginning when I installed RHL9. It appeared later on. Still, I wish to know what is causing it so that it won't appear the next time I install Linux.

-Ol

If it's a hardware problem, dmesg and looking at /var/log/messages will give you an idea. You might also think about installed lm sensors so that it gives you temperature reports of CPU and whatnot.

-twantrd

And check that the CPU fan is not clogged with dust. LM sensors should also be able to report fan speed, check that it is actually working.

I don't think my CPU is overheating. I leave my PC up and running 24/7 but this problem arrived just recently - some time after installing Linux. Earlier, when I still had Win98 installed, it was running UT fine without freezing up. I believe this would be more of a technical problem rather than a problem of neglect for I recently cleaned the CPU casing and the fan. It now runs as quiet as a mouse.

Have you taken a look at the logs to see if anything might be related to hardware issues?

-twantrd

I have but all I see there is the entire boot log and after that a couple of NIC log entries if I remember correctly. The file itself is too big for kWrite to open. Is there anything specific I should look for in there?
Sometimes it even freezes when I am looking at the logs in kWrite

Hmm, use tail -50000 or use more|less|vi.

-twantrd

Okay. I downloaded the file to my windows box where I can open it without a problem. I can see the boot log in there but after that there are a lot of those NIC entries that look like this:

And it goes on like this forever. However there are some time gaps in the log which is when I believe the PC stopped responding. The time gaps appear as follows (including the boot log):

Now this basically doesn't provide much info on why the PC stops responding as it doesn't log any errors. I do not believe it's a hardware error because it started to occur recently. It used to run fine before. I tested the RAM modules with GoldMemory and they passed. Would it be possible that my firewall has something to do with it? Although I'm not sure about that because the Flash also caused it to freeze. It used to freeze before I installed Flash player so I doubt Flash is the problem. I have 3 different kernel versions and it does not matter which one I choose because it freezes eitherway. Either I'm just unlucky with my linux choice or there's something terribly wrong with the PC.

Bummer, that log doesn't show any relevant error to your problem. My next step in debugging this would be to run sar (System activity reporter) every 5 mins via cron. The next time the box dies, at least we'll be able to see what the load was and other useful information.

-twantrd

And how to do that?
P.S. I don't believe I have sar installed... "whereis sar" does not return any results.
Man I can't even work on that PC - it freezes like every 10 minutes while I work on it.

It only appears to freeze on high CPU loads which is exactly when I work on it like browsing internet, opening logs, etc. That's why it freezes so often when I use it and rarely when I don't.

Curious. You said I'd need to have sar run every 5 min. So in order to get the most precise measurement, will the PC have to freeze a second after sar is used?

sar comes with the systat package. Install that and you have 'sar'.

To actually get the most precise measurement you can run sar every minute. However, since you believe that the machine just freezes when there is high cpu load, do a cpu burn test and see. You can find simple shell scripts out there on the inet that will help you achieve this. Another option is to swap out the cpu and try loading all your programs again.

This particular situation you're in would require a lot of debugging time and patience. If you don't have the time nor the resources, why not build another linux box on different hardware and run your apps on that instead?

-twantrd

I'm going to do a simpler test that will show if it is an actual hardware bug or software bug - format and install a new linux distro. If it continues to crash then it must be hardware. Honestly I was just curious if there was an easy way to fix this crashing but since I'm at home only on weekends and I barely have any time for myself, I'll have to spend that time wisely. Debugging would take too much time since I'm relatively new linux user and I have to learn things before actually attempting them. I will install a new distro and if it continues to crash I'll do what you suggested.
Thank you for your help.