[patomas]

Hi people

I have a little problem since a month ago.

If i am in any kind of visual environment, fluxbox, kde, etc, the system, suddenly, can crash. There is no previous signals or any kind of symptom.

If i am in the console, before startx or after exit the environment, the system works perfect for as long as it is required.

When the problem started, i thought it was caused by a buggy ati-driver, so i uninstalled it and went back to the old version, but the problem continued. Since I upgraded many things the same day and since I wanted to test some other things, I decided to reinstall a new system, so Ii did 2 days ago.

But the problem is still here.

There is no information in the logs, at least any thing i can find.

When this situation happens, the machine needs a hard reset (4 seconds power off), and then it is possible to start again.

So, the question is, do you have any suggestion to test specific parts of the hardware to know what is wrong? I think the problem is in the graphic card (ati x1300 pro). But fglxgears and similar work normal, i mean, they do not crash the system and print normal and expected information.

Any way, if you want me to post information, i have a whole set already in files, lspci, lshw, dmesg, etc, just mention and i will post it.

Thanks for the time and patience.

Bye

[aliov]

Post the content of xorg.conf wish is in /etc/X11, by the way it's a 64 bit machine ?

[i92guboj]

Originally Posted by patomas

If i am in any kind of visual environment, fluxbox, kde, etc, the system, suddenly, can crash. There is no previous signals or any kind of symptom.

Define "crash".

Does it halt? Is turned off? Does it hard-reset without any warning? Does linux do anything? Does the screen stays black? Or maybe it hard locks and you have to reset it by hand?

Describe de symptoms as accurately as you can.

[juicydata]

How's your heat / power? Do you have onboard video you can fall back on?

Try removing the ati x1300 pro, and running without (just switch the Driver line in xorg.conf to whatever chipset your mb supplies).

You may be able to monitor some temps in /proc/acpi and it's children.. but I'm not sure if you'll get goodness there.

If you're using the prop drivers, try switching to the open ones. Only bother with this if your pre-install and post-install both used the same drivers.

Just some ideas, best of luck.
-juicydata

[patomas]

Hi

First, thanks for the answers and the help, and sorry for the delay in my reappearance here.

Yes it is a 64 machine.

About the xorg. well, i can post it, but i think it is not going to be necessary right now.

About the crash, well, as i mentioned in the original message, the computer needs the hard reset to be turned off and on again.

About video on board, no, i don't. The motherboard is an Intel 975XBX.

About the temperatures, i use to monitor them with lm_sensors and gkrellm, and from time to time, confirm the readings with the BIOS readings, and they where accurate.

Well, after a few days of testing, the things seem to be more stable.

This is what i have done and tested.

In the same computer:

1. ati-drivers (problem reported here)
2. radeonhd (same problem, less frequent)
3. vesa (no problem at all)
4. memtest (all ok)

So, the drivers with more options, and more demanding causes more problems.

I haven't used cpuburn to test, but since my system use to run boinc the whole day, I think that is quite a test. This computer is 1 year old, more or less, and has been running boinc for almost the whole year.

From another computer when the system dies.

1. Ping to this one and it was working. ( Test made when the problem happened again ).
2. Connected through ssh and checked logs; no problem at all reported in any of them.
3. Connected through ssh and checked about the X being loaded and working. They were not in memory. Neither fluxbox

So, from the other computer i did reboots or halts when necessary.

With those tests i confirmed that the system doesn't really dies, just the graphical system.

The next test i did, was to connect a fan close to the graphic card and in front of the southbridge and then the system seems to be more stable. Obviously this point out to be a temperature problem, but the thing is the system did not report any excessive heat, not in the core, not in the bridges. And the graphic card was never hot enough that i couldn't touch it.

Any way. The system seems to be more stable now. Only one problem starting the X system today morning, but again, no logs. This time i could not connect from another computer, so i had to reset manually. May be it was related with the original problem, may be not

Thanks again for the help and the answers, I will keep doing some more tests until I'm convinced that the system is really stable, but if you have any theory or idea, I'll be glad to hear it.

Bye