Random Hard Freeze - Driving Me Nuts!
 

I'm running linux... supposedly super-stable, but ever since i upgraded my hardware a year ago, i've had the persistent problem of getting hard freezes. Therefore, i know it's a hardware issue.

The symptoms are: suddenly and without warning, the screen locks up completely, cursor stops dead in it's tracks, all effects stop, sounds stops, keyboard stops responding. Full Stop. It does this at seemingly random times. If it does this when the screen is in sleep mode, it will not wake up.

It's exihbited this behaviour with Mandrake 9.1 and 10.0 and Debian. It happens on the installation and in normal desktop use. However, i think it happens more when the HD is in heavy use.

So... i bought a new hard drive and complete reinstalled everything. No help. i tried swapping out the RAM modules one at a time. No help. I reset my MoBo settings to FAILSAFE and it seems to have helped a little but i still get the lockups every day or two.

My next step is to get a different motherboard. Do you think this might help? I'm running out of options. What would cause the system to lock up so suddenly? It's driving me nuts

I'm currently running an FIC AU13 w/ nVidia nForce2, AthlonXP Barton 2500+, 2x 256MB DDR (not sure about timings, but lower end), nVidia GF4 MX 420 video. HD is a brand new Western Digital 80GB w/ 8MB buffer, replacing an equivelent Maxtor.

If you can offer a suggestion, i'd really appreciate your input. Thanks.

Based on your description, I'd say it's either bad RAM, or a bad controller. Do you have mis-matched RAM in your box (different speeds, brands, etc)? If so, those differences could cause an incompatibility and freeze up the system. If you are using only a single stick or a pair of identical sticks, then it may be that one or both of the sticks has a fault. I would suggest running memtest86 on your system. memtest86 runs a set of diagnostic tests on your RAM, and should uncover any problems if there are any. If you've got bad RAM, you'll need to replace it.

Ditto on the bad controller. There is really very little that you can do about that other than to replace the mobo.

I would suggest that if the problem(s) started happening immediately after the upgrade, then something that was done during that operation is responsible for your system freezes. I'll take a guess though and say it's your RAM, particularly if you yourself indicate that your RAM is "lower end". You pretty much get what you pay for with RAM, and I'd recommend going with a quality name-brand (such as Kingston) rather than a no-name, generic brand. Yes, it will cost a bit more, but it's worth it. Good luck with it -- J.W.

Hehe - I got about the same hardware and the same problem.Even put the exactly same drive in my box to replace the maybe bad one but it already froze up again.Me starts suspecting the 2.6 kernel and/or the nvidia driver.I'll report back if another kernel solves the problem.

I ran memtest for 48 hours straight through 9 passes with 0 errors. As i mentioned, i've tried testing each stick individually and they lock up either way in any configuration. It's Kingston RAM. I've had bad RAM before so i bought the name brand this last time. it appears to be fine. It's two sticks of identical RAM.

I'm running kernel 2.4.25 for Debian, but i've had the same problem with 2.4.21 with mandrake 9.1. I don't think it's a kernel bug. I'm running out of things to blame, so my finger is now on the motherboard. Any reason why i *shouldn't* get a new one?

That's a bummer, and the outlook doesn't look good. Nobody likes to have to spend the money, but if you don't think it's the memory, and you don't think it's the kernel, and it's not due to any apps that you are running, then I would say that it may very well be a bad controller or other component of the mobo. Certainly if anyone else was facing the same conditions that you describe, no one would fault them for replacing the mobo. Not that it really matters, but if I were in your shoes, I might make a few last ditch attempts to see if it was salvagable (maybe by unplugging and reseating every cable and plug) but if the freeze-up happened again, I'd throw in the towel and just replace it. Good luck with it either way -- J.W.

I'm currently running an FIC AU13 w/ nVidia nForce2, AthlonXP Barton 2500+, 2x 256MB DDR (not sure about timings, but lower end), nVidia GF4 MX 420 video. HD is a brand new Western Digital 80GB w/ 8MB buffer, replacing an equivelent Maxtor.

I had lockup problems with my ASUS A7N8X Deluxe (nForce2) motherboard which were minimized (but not completely eliminated) by disabling APIC. Try adding "noapic", "nolapic", "acpi=off" on your bootloader params. Other people were able to fix this lockup problem with nForce2 boards by flashing their BIOS to the latest ones.

Better still, get an Intel based motherboard... no lockups on an MSI and an Intel board I am using.

cyclo

I put a new 2.4 kernel on the thing and reinstalled the nvidia driver for the graphics and it is humming along fine ever since.Dunno for sure but my money is on the nvidia driver creating problems with a 2.6 kernel.

I THINK I MAY HAVE SOLVED THE PROBLEM!

A lot of websites and postings have remarked that there is a bug regarding nForce2 boards and APIC (not APCI). I disabled it on the grub boot line with:

kernel /boot/vmlinuz-2.4.25-1-386 root=/dev/hda9 noapic apic=no APIC=no nolapic ro

i'm not sure which one did the trick, so i did them all :-) my computer has been running for 3 or 4 days nonstop and i've installed nearly 100megs of software onto the hardrive to test it out, and no crashes!

If you are experiencing hard freezes, try disabling APIC and see. Make you you disable it in linux because i turned it off in my bios and linux message file still says it enabled it. (until i disabled it in linux as well)

leiavoia - that's excellent news, and if it's been running trouble free for 3 or 4 days, then it sounds like you've solved the problem. Congrats on that, and thanks also for posting back with the solution. That is always a huge help for future readers who may be having this same problem and find this thread. -- J.W.

Kernel 2.6.7 supposedly also fixes those issues - don't have them so I can't check.To keep you going it crashes with the 6106 nvidia drivers if usb2 is compiled in the kernel:rolleyes: :rolleyes: