Any aid, including indicating if the thread is at the wrong place, is welcome! Thanks!

Hardware:
MoBo: Asrock 770 Extreme3
CPU: AMD Athlon II
RAM: 2 gb DDR3 (single stick, single channel mode)
!!!* TESTED PASS: 7 ERROR: 0 *!!! with memtest 4.x
VGA: Gainward Bliss nvidia 9600 GT 512mb

this is new hardware.
I have two hard drives, hdc (ide) and sda (sata).

I loaded default bios settings.

I used openSUSE-11.4-NET-x86_64.iso from usb stick according to http://en.opensuse.org/SDB:Live_USB_stick and choose KDE.

Installation didn't work out first and second time. It started but after random time two led blinked on keyboard (kernel panic?).

After that I started to use the failsafe kernel of the installer. I used the same ext4 partition, but formatting was on.

I DEselected the office, multimedia and game software package groups.

This time the installer succeeded and right after that I got to desktop.

I added manually the nvidia repository for 11.4 with the key
ftp://download.nvidia.com/opensuse/11.4/
and installed some ..G02.. packages what seemed to be the needed ones, based on an older docs for 10.x what currently I cannot find. (I didn't found the newer docs with the ymp scripts first.)

I hit Ctrl+Alt+Backspace twice, the screen went blank and never got back.
I pushed the power button and it _instantly_ turned off. (???)

I booted desktop suse, and I got a bunch of segmentation fault messages, and many "failed" texts...

It stopped and never got to login.

I tried the failsafe, it worked, no segfaults.

There was text login, i logged in and did
sudo nvidia-xconfig
what I forgot earlier. I typed
sudo startx
(simply startx didn't work due to file permissions on the log, i guess it's normal)

X started and KDE showed up with that info box for suse.
I've heard the fan for nvidia vga card slowing down since until it's propertiary driver is loaded it works always on max speed.

After that everything freeze, BUT the mouse worked.

I hit Ctrl+Alt+F1, the screen went blank and the monitor turned off.

I hit Ctrl+Alt+F7 and I got a blank black screen with a mouse pointer that I still could move.

I hit the power button in both modes but nothing happened.

After a while I turned off the PSU.

I really would use some help... it seems a hardware error, but the RAM seems fine. I'm currently running more passes after that I want to check the cpu temperature. However it never turned off after hours of usage (memtest) so I guess it will be ok?

I've also tried an old linux install what worked but seemed to crash randomly.

I've even tried windows xp setup but it's also crashing randomly in it's gui part.

Segfaults are a sign of madness, so it's just a matter of deciding what's mad. You checked the memory.

When the problem started, what had you done last? That's usually it.
Installations not working sound like dodgy media (install cds). You can simply trash that install, and try again. This time stop the install on the first error.

Installed the nvidia prop. driver. Maybe I should try without it. However all installers also fail randomly even without it.
I forgot to mention that I'm using USB sticks. (I don't have my optical drive plugged in due to the geometry of my case.) Better check them though.
I can't actually stop at the first error because when the installer freeze there is no error message and I can't do anything (I don't have control neither with keyboard nor with the mouse). The keyboard blinks, it seems like a kernel panic to me, but I can't change to check what is it.


Now I'm testing the Mobo and the CPU.

EDIT: I thought everything was fine after adding apm=off edd=off, but actually it just took some time until segfaults happened!

Now my OS usually boots but after half an hour or so (I used firefox, installed packages and occasionally played tux racer) segfault things appear, and the kernel is sending me messages. I wanted to log them and post here but by the time firefox crashed instantly whenever I started it, and when I tried to find a simple text editor, the whole system just freezed/went down (can't remember).

CPU/MOBO diagnostics
I ran cpu/mobo diagnostic tool "PC-Doctor" and "PC diagnostics" from hiren's bootcd 10. However it couldn't test both cores due to "I'm running a memory manager" (I guess the bootcd is), ALL other test PASSED both on mobo and cpu.

All but the cpu stress test in this application, what instantly failed, but I think this is due to it trying to test the 2nd core in an instant (if I got it right based on the screen).

I ran other CPU stress test with "prime95" and cpu tests "S&M" test from hiren's bootcd 14 and it all PASSED.

Hiren's bootcd 14 has miniXp and it crashed several times when I was doing several other things, however.
Most notably it crashed/lagged heavily every time when I tried to start utorrent (NOT included in the bootcd!) and it never succeed. Other thing to note that it freezed twice when trying to do something with realtec PCIE something (for enabling internet access in the miniXP). On the other hand, other times it succeded.

I found the log of my last kernel crash.
It was around "Aug 25 23:". (22:59:xx segfault)
I also attach the lspci -vvv output.

Ran mprime, attached screenshot, output log here:
http://pastebin.com/9kQSakYq

Attached Thumbnails

 

Attached Files

	messages.log (162.0 KB, 20 views)
	lspci_vvv.log (32.9 KB, 21 views)

Take care to observe the pattern of misbehaviour. You only can do this. When did the first error happen? It strikes me (from post #1) that it was here:Try a live dvd or cdrom; no need to install optical drives, open the case and support it somewhere. Observe closely the voltages. 5V cannot be below 4.8, and I wouldn't like to see 3.3V below 3.2V, although it may work on that. google your cpu and check that the correct voltage was automatically selected

Thanks for the reply and info!

I'll try the real DVD install.

Measuring voltages, you mean physically with a multimeter? (My case doesn't have a temp display.)

I attached some screenshot from my current installation.

One thing I am not happy about (apart from the random segfaults) is that my
w83627dhg-hwmon1/temp3
shows 55 to 145 C! (Note: not constant.)
On the one hand it's brutally high, on the other hand it seems so brutally high that it doesn't seem real. Wouldn't the chip/IC die in no time and never turn on again with a temp like this?
I don't know what temp3 is. All other temps are normal.

Bios health_of_system checks are often the best. If you haven'g got them, by all means use a multimeter if you have such handy. But be careful, as many a circuit has been destroyed by a slip of the probe.

I think you're going in the wrong direction with all this health test stuff, btw. Reread my previous posts and you'll see I'm not prepared to go past the funny installation as a start of the trouble.

Your story has been "I installed this software, and had that trouble; then I installed that software and had this other trouble." Might it not be an idea to get a clean flawless install going?

Yes of course, thank you. I'll try it ASAP.

By the way, I did these measures 1) before I read your post 2) because it was tempting to do anything with my only working (even if only working for <1h) system in a week.
(The time of my posts are not too important since I don't have the luxury to always finish a post if I start to write it .)

ROTFL.
Go to it at your leisure and report back.