A hard lockup

***glennzo*** · #1 8th July 2012, 10:57 AM

In the image I intend on attaching you will see a photo of my desktop monitor/screen when said desktop locks up hard. It is a seemingly random lockup that doesn't seem to discriminate against what distribution I'm running. I have the following installed and the system has locked up when running all of the Linux OS's. (no matter what OS I'm running)

Fedora 17
Fedora 16
Arch Linux
Linux Mint 13
Windows 7

The desktop system stats are in my signature. In all cases I'm using the default nouveau graphics drivers. In the case of Windows 7, I rarely boot that so I don't remember if it happens with that OS or not. So where would one begin with troubleshooting? Would there be any clues in the log files, like something that occured just prior to the lockup?

marvin_ita · #2 8th July 2012, 11:43 AM

The fact that it happens with all OSs make me think that is a hardware related problem.

First, try to run the system with pc case open to see if the problem is temperature related.
If not, you can try running the system without some RAM (if you have more than one module, leave only one installed to check if it's a faulty module). Or if you have another video card try to replace it.

jpollard · #3 8th July 2012, 12:15 PM

I get that display occasionally too - but only as a background image. The foreground has the usual/normal GDM login. Variations on the background have happened as well, but most of them seem to be an improperly resolved image from the last desktop background image.

I suspect a timing error (not with the hardware) in resetting the display (due to logout/switching user) and the background image loading.

A synchronization error with multiple processes doing things to the display.

You might try a "ctrl-alt-backspace" and see if the reset clears things. Alternatively, try ctrl-alt-F2 to bring up a console window. If it switches to the console window then the X server is working normally.

I'm using F16.

It isn't hardware.

***Dan*** · #4 8th July 2012, 01:05 PM

And an opposing opinion ...

You say hard lockup. Assuming that means no keyboard, mouse or other means of interacting with the machine are functioning, I've been down this road (or one very similar) before, so I suspect it is hardware. In my case (twice as a matter of fact) it turned out to be the video card going south.

http://www.forums.fedoraforum.org/sh...d.php?t=233675
http://forums.fedoraforum.org/showthread.php?p=1354287

So my first order of business would be to turn the box turtle, and have a close look at all the capacitors on the motherboard and extension cards, with a particularly close inspection of the video card.

#5 8th July 2012, 01:56 PM

Quote:

Originally Posted by Dan

And an opposing opinion ...

You say hard lockup. Assuming that means no keyboard, mouse or other means of interacting with the machine are functioning, I've been down this road (or one very similar) before, so I suspect it is hardware. In my case (twice as a matter of fact) it turned out to be the video card going south.

http://www.forums.fedoraforum.org/sh...d.php?t=233675
http://forums.fedoraforum.org/showthread.php?p=1354287

So my first order of business would be to turn the box turtle, and have a close look at all the capacitors on the motherboard and extension cards, with a particularly close inspection of the video card.

+1 for a video card issue.

nonamedotc · #6 8th July 2012, 02:39 PM

A friend of mine had a similar issue recently. It turned out to be a faulty video card. It was not the same video card though. If it had been a laptop, my first suspicion is would be the led cable - may be a faulty connection?

***glennzo*** · #7 8th July 2012, 10:55 PM

Hard lockup. No keyboard, no nothing. The video card is onboard so I'll have to look and see if I have a replacement handy.

This happens after the system has been up for minutes or hours, not at login. As I type I'm running Arch. Been up over a day so far without issue.

PHP Code:


			
3.4.4-2-ARCH
 17:56:22 up 1 day,  2:49,  2 users,  load average: 0.72, 1.14, 1.40
USER     TTY        LOGIN@   IDLE   JCPU   PCPU WHAT
glenn    tty7      Sat15   26:49m 24:08   0.06s gdm-session-worker [pam/gdm-password]
glenn    pts/0     05:06    0.00s  0.06s  0.00s w
glenn    tty7         2012-07-07 15:14 (:0)

glenn pts/0 2012-07-08 05:06 (:0)

george_toolan · #8 9th July 2012, 11:14 AM

Can you still ping your computer from your notebook?

If you enable Magic SysRq key you should at least be able to safely reboot after a lockup ;-)

***Gareth Jones*** · #9 9th July 2012, 04:22 PM

Much as I'd hate to compound your agony, I'd suggest running Windows 7 for a while, just to see if a completely different graphics stack works. If it does, that doesn't rule out hardware problems of course, but if it fails then you can be pretty certain that it's flaky hardware.

***glennzo*** · #10 10th July 2012, 11:39 AM

Well, since I've posted this question I've had one lockup. Oddly enough, I had rebooted the computer and walked out of the room. Boot defaults to Fedora 17. When I returned I was greeted by the same garbled screen and a hard locked computer. Didn't even have a chance to log into the desktop. I think that this throws a wrench into the works, thinking about jpollard's comment that "it isn't hardware". This behavior, I believe, is exactly what he described.

I think it is hardware related since it happens on more than one OS / distribution but alas, I haven't taken or had the time to remove the cover and begin real diagnostics.

As you can see from the user agent icons, I've booted into Windows 7. Since there are only a few minutes left before I have to leave for work, I figured I'd boot Windows and leave the thing running while I'm away. We'll see what things look like when I get home.

I appreciate all the comments and the interest. Thank you.

Iron_Mike · #11 10th July 2012, 09:26 PM

Glenn, do a test with memtest and see if memory chip is getting ready to flake out on you.

***glennzo*** · #12 11th July 2012, 07:58 AM

Quote:

Originally Posted by Iron_Mike

Glenn, do a test with memtest and see if memory chip is getting ready to flake out on you.

Done. Ran memtest overnight (about 7 hours). Several passes were made without error. I'll run it again when I leave for work today so it will run all day.

Almost forgot. I had Windows 7 running all day yesterday. When I got home I shut down Windows (reboot) and while that process was executing the computer locked up, giving me that same display shown in the screen shot. That proves that this problem is OS independent and hardware related.

marvin_ita · #13 11th July 2012, 05:29 PM

I had similar lockups in the past, but instead of a garbled image I had the Fedora logo displayed with the system completly freezed. I also did several memtest, but the RAM always passed. BUT since I've changed the RAM I never had another lockup.

However in my case I always had those lockups when the temperature was high and system under stress (with a supplemental fan directly over the RAM modules the system worked well). In your case, looking at the garbled image, I think the problem is probably the embedded video card; I would suggest to remove one of the chassis fan and put it over the video chipset and see what happens...

***glennzo*** · #14 11th July 2012, 05:41 PM

Hello marvin_ita. That's' a good suggestion. One thing I do have plenty of is spare fans. I'll give it a shot.

I do wonder though, if the video chip is bad I know I can install another in the PCI Express slot and disable the on-board video, but would disabling the on-board video chip keep it from getting too hot and causing issues?

marvin_ita · #15 13th July 2012, 08:22 PM

(sorry for the late reply)
Good question, I would say yes.
At least it will reduce the onboard GPU frequency to minimum (if it doesn't turn it off at all), so it will not be so hot. And theoretically the OS will not see it at all, so the system shouldn't be tampered by the onboard video problems.