SUSE 11.2 "hangs" randomly with Dell Dimension Desktop

I love openSUSE (using 11.2). I have it installed on a Dell Dimension 2400 (2.4G Celeron CPU, 2G RAM), but I have a huge problem. It will randomly freeze during operation. I’ll be able to move the pointer with the mouse but unable to click anything and I have to hold the power button down to shut it down and restart. The problem seems independent of the program I am using or what I am doing with the machine. I have not been able to recreate the problem. One more detail. The same problem occured when I had Ubuntu 9.10 installed on the same machine.

Any ideas?

It will randomly freeze during operation

I have not been able to recreate the problem

I’m not quite sure what to make from that
But when you say randomly, do you mean randomly once a day, or once every 10min?

If it’s frequently random, you could leave a terminal open running ‘top’ and see if you can catch any obvious indications from that.
Check .xsession-errors too

It’s possibly the Intel graphics. I had a similar problem with my Compaq/HP laptop, but typically only when I was doing graphic-intensive work (running Celestia or Google Earth or editing graphics, for example). I haven’t had the problem on my laptop since I upgraded to OpenSuse 11.3 and changed to Gnome. Maybe the drivers have been improved.

You could take a look into /var/log/messages, it often provides info about such errors.

Good info smpoole7 and gropiuskalle. caf4926, I too thought I was a little vague there. By “randomly” I meant that the problem has no noticeable pattern. First I thought it would happen within 10 minutes of booting, but no I had it running a few programs (firefox, bluefish, the terminal) for hours without an issue. Then, when my daughter played an online flash game it froze. Upon reboot I opened firefox, and it froze in 5 minutes (no flash involved that time). I began, of course, to assume it was firefox, so I kept it running without using firefox and my wife found that the Kpatience game she was playing was frozen.

A Dell Dimension 2400 is pretty old. I hope to replace my mother’s Dell Dimension 2100 in a couple of weeks (although her Dell is a lot slower : 900MHz CPU and 512MB RAM).

You could run a memory test and let it run over night to see if there may be a hardware memory problem.

The OP quotes a somewhat improved spec than your Mom’s old box Lee.

OP: Please make sure you don’t use desktop effects and avoid full screen flash. Take note of the other comments posted re: identifying the cause

Indeed !! In fact I stated so myself. But lets not get lost on that and ignore the suggestion I made.

Freezes can be difficult to track down. My experience in general is a freeze (that can not be easily pointed to software) are due to (in no particular order):
(1) bad graphic driver or graphic hardware
(2) bad ethernet driver or ethernet card hardware
(3) bad memory

A surf indicates the Dimension 2400 has a version of Intel Integrated grapics:

and a further surf suggested it is the Intel 865G which I think is the Intel GMA 3000 ? I’m forced to guess/assess via a surf here, because this important information was not provided. It would be nice to have it provided.

One could surf to see if the Intel driver associated with the Intel GMA 3000 has any problems with the current kernel.

But we need specifics here, and not generalities. In the mean time, I still think it a valid and good suggestion to run a memory test overnight while everyone is sleeping. It does not hurt and it may provide useful information.

I should have ‘amplified’ (2) or modified it with a (4) bad wireless driver or wireless hardware (in addition to bad ethernet driver or ethernet card hardware).

apriest, do NOT forget to followup gropiuskalle’s suggestion:

… after such a freeze, you could also boot to a liveCD and in addition to looking at the /var/log/messages, copy the /home/user/.xsession-errors file and also the /var/log/Xorg.0.log file that are on your HARD DRIVE (and not the one’s in RAM from booting to the liveCD). Booting to a liveCD is necessary is to get the version of those files BEFORE the liveCD boot. …

Fortunately /var/log/messages is continuous and not over written by a reboot (but rather is added to).

The “freeze” occurred again just a minute ago as I was responding to this message. I’m on the live CD now trying oldcpu’s suggestion.

What should I do with the /var/log/Xorg.0.log file? I’m not sure what I am looking for.

Also, if it helps, the chipset is Intel i845.

Oooookay. I get a permission denied when I try /var/log/messages as su!

You may be in the realms of complicated here, I don’t know.
But you need to mount the / partition containing your installation. Personally, I always use parted magic,because there is never any problem: mounting, reading, deleting…you name it…
Note@oldcpu’s comments: (and not the one’s in RAM from booting to the liveCD)

Some of the i8xx series have problems with the Linux kernel. I note the i855GM has the worst in this thread: Intel GPU 8xx issues, will 11.3 have them too?!

But the i845 series has some problems. You could try switching to using a different graphic driver … ie try the intellegacy driver (if you have not done so already). Do you need advice as to how to do that?

Hmmm … you are using openSUSE-11.2 … I can’t recall if the intellegacy was available for 11.2.

I was able to mount the / partition while using the live cd and copy the /var/log/Xorg.0.log to a jump drive (just for simplicity). However, I did not know how to use that file to get at my problem.

Another update: I did do a memory test. It reported a Pass of 100% but also listed two failing addresses. What can I conclude from that?

There will likely be more useful information in the /var/log/messages. Look there first.

What version of the Intel graphic driver are you using? I suppose it is possible there have been patches to that chipset. It is difficult to track this (at least I find it so).

I’ll try help you determine this driver version. What is the output of:

rpm -qa '*driver*'

with that information, one can then surf and find out.

For example, if you have the xorg-x11-driver-video-7.4-87.91.1 then one can download for inspection but NOT install the source (.src) file from the source repository for updates: 

and then look inside it with:

oldcpu@hal2009:~/rpms> rpm -qip xorg-x11-driver-video-7.4-87.91.1.src.rpm -l 

which indicates:


ie openSUSE-11.2 update has the 2.9.1 Intel driver (with patches).
One can then search on the 2.9.1 intel driver (and the 4 patches intel-G33-1mb, intel-NoFBC-945GME, intel-bfo17988 and intel-buildfix) to see what updates are present, compared to what is now available for openSUSE-11.3.

In the case of openSUSE-11.3 one can go to the source directory for its rpms:

and download for inspection (but not install) xorg-x11-driver-video-7.5-15.2.src.rpm, and xorg-x11-driver-video-intel-legacy-2.9.1-1.9.src.rpm and then check what driver version and patches are inside:

rpm -qip xorg-x11-driver-video-7.5-15.2.src.rpm -l 

which indicates:


i.e. openSUSE-11.2 “intel” driver has version 2.12 of the intel driver with the patches intel-buildfix, intel-copy-fb, and intel-mbp_backlight.

and for the intellegacy driver:

rpm -qip xorg-x11-driver-video-intel-legacy-2.9.1-1.9.src.rpm -l

which gives:


ie openSUSE-11.3 with the intellegacy driver is also using the older 2.9.1 driver, but it has a MANY extra patches that do NOT come with openSUSE-11.2’s version of the 2.9.1 Intel driver.

To see if any of those patches address an i845 problem, one would have to research the patches.

Plus there could be kernel updates that help the i845 between openSUSE-11.2’s 2.6.31 kernel and openSUSE-11.3’s 2.6.34 kernel. That also would take time to check.

Two more hints on the logfiles: as oldcpu mentioned, /var/log/messages is chronologic, so you should simply look for the exact time of your last freeze. The lines in /var/log/Xorg.0.log are prefixed (for example (WW) = warnings, (EE) = errors), so it’s easy to grep lines that may show you the culprit:

grep '(EE)' /var/log/Xorg.0.log

If a memory test fails you have flaky RAM. One or more locations are not reliable. This definitely can cause random problems.

So, eventhough the Memory Test said 100% Pass, those two failed addresses are enough to cause a freeze.

I never mentioned that I am dual booting this machine with WinXP that works without problems.