Linux Lock-ups?

In my experience, Linux has been a pretty stable platform (i've been able to leave my machine on for days), but I have noticed that on a few different occasions it has locked up for some unknown reason and I am forced to do a hard reset.

I don't know if I am doing anything wrong here, and I am not sure whether there is a more graceful way to get out of system lockups than what I have been doing previously. Any help is really appreciated.

Also, i did read an article about some interupt key patterns to use if the system locks up, but this last time even that didn't work (neither <ctrl><alt> backspace nor <alt> <prt scr> REISUB worked for me).

I wrote this post because I found it pretty disturbing that all i was doing was watching a couple of videos on youtube and the system just ground to a halt. I sincerely hope this is a stupid configuration error on my part as opposed to some sort of hardware malady, because as I understand it, Linux will only do this when something has gone horribly wrong. On a related note, there may very well be a problem with the linux version of the firefox flash plugin since it has caused firefox to lock up several times ( but never system wide before).

Thanks guys.

-Zorak

Flash player is notoriously unstable and will lock up all the time.

You are correct however that a full system wide lock up could potentially be something far more serious.

Well, knightrous, on a couple of occasions i have had the lockup with music still playing in the background (and it could very well be the nvidia driver as i have an nvidia card myself) but this last time, the music from the youtube video kept replaying about the last 20 milliseconds again and again and the entire system locked up and wouldn't respond to any of the interrupt combinations i know.

Is there anything I could do to check my system aside from just running memtest86+ to check for ram problems (i.e. any other good diagnostic software)?

Thanks

-Zorak

I know this isn't a real benchmark, but on a couple of occasions after this lockup I installed and played the Crysis single player demo for several hours (its fun messing around) and I didn't experience any crashes due to overheating or the like while I did that. Additionally I have been able to do the same with no ill effect in the Unreal Tournament 3 demo. I would think that since windows is less stable over all, if I had a loose heatsink or too much thermal compound, i'd have the same issues there, but I don't. I guess as soon as I get the chance I will try running a few of the benches that you suggested ( i have like 3 tests and 2 projects coming up )

Next time I experience a lockup in Linux (which i hope doesn't happen soon), I will try pinging my machine from another box and maybe i can gracefully kill X remotely since I can't do it via <ctrl><alt> backspace or via a virtual terminal.

also, about compiling the kernel, would i just grab a copy from kernel.org, untar it to a temporary spot and just type make? I just want to make sure that I won't inadvertently replace my current kernel with a new "hand-compiled" kernel because I have no experience compiling kernels and I don't want to have to spend time fixing my machine when I really need it for an upcoming assignment.

Also, an a somewhat unrelated note, audacious media player just broke on me last night for no reason! I am hoping that it was just a bug in an update or something because it is strange when you have a large playlist of files with random select and repeat options enabled and it "randomly" elects to play the same song again and again...

Anyways, thanks guys.

-Zorak

To compile the kernel just grab the latest full source from kernel.org currently http://kernel.org/pub/linux/kernel [...] .1.tar.bz2


F = Full Source

extract it

make menuconfig

use tab to exit from the configuration menu and let it save the config when prompted

make

sensors

make clean

make

sensors

rinse and repeat


you need to install and configure sensors if you don't already have'em



here's some benchmarks to google:

mysql-bench

postgresql

tscp

POV-Ray

lame-mp3

gzip

bzip2

oggenc

http://lbs.sourceforge.net/


GL

That's not a complete lockup, that's X locking up. Keyboard, mouse, and video are controlled by X, so the music still playing tells me it's not the whole machine locking up. If you have a second machine, try sshing into the "locked" one. Chances are you can do this as you would normally. Kill the X process, rmmod the GPU driver (nvidia), modprobe the driver, restart xdm/gdm/kdm and you're good to go.

I have had this happen on my machine when playing ETQW and the aforementioned steps work well. Just beware that if you restart X without rmmoding and modprobing the GPU driver and X locked up because of the GPU driver, you actually will lock up the system (your ssh link will die.)

Check to see that you are putting enough volts to your RAM in the BIOS and making sure that it is running at the appropriate speed and timings. If you are OCing, then return the clocks and volts to normal.