SEQ Freezing

[Amadeus]

I've noticed for quite some time that SEQ (and the entire Linux OS) will "freeze" up after enough time in a particular zone.

I was wondering (ie, Zaphod!) if there was a particular syntax you would want from gdb using core dump that might help in debugging this.

Thanks!

[Dedpoet]

The OS shouldn't be freezing if an application crashes, that's the nature of the OS. If you're whole machine is locking up, it's most likely a hardware problem.

I know you've been around here a while, so I assume you've been running Seq for quite some time. Did your hardware change? Are you doing anything else differently?

[Amadeus]

No...this started happening ..hmmm...it was probably around the last MAJOR seq fix that we had a few months ago.

It is a complete computer freeze though ....the only thing that's quirky is that the keyboard lights are flashing every time ;)


I notice that it happens almost every time if I forget to close SEQ after having turned off or logged out of EQ proper on my windows machine.

[Freakyuno]

Sounds like an isolated problem. I've kept up to date with all the releases, and my current SEQ session has been running for atleast a week solid, without ever beeing shut down and restarted.

Have you tried to do a makeclean and recompile / install?
Keyboad lights flashing means your bios is trying, or has put
your computer into some type of suspend, or sleep mode 99% of the time.

[Amadeus]

hmmmm...I suppose that could be the case. Where is the freakin setting for sleep mode or whatever in XWindows?

I only use xwindows for SEQ (I'm a console man) so it's always new for me ;)


I'll also check out a crash dump next time it happens to see what's up.

[fryfrog]

also, while the system is frozen... can you ssh into it from another computer? my dual celeron system used to be quite unstable, but after i pulled it out and sucked about 4 cats worth of hair and 3 dust bunnies from it with the vacuum cleaner... its been running a lot better.

the procs are overclocked 366@550, so SOMETIMES when i am compiling large things (gentoo), it will tank... but this is quite rare and not enough to really make me wanna go down in clock speed. usually i just init 3 / stop x before i compile anything.



xset s off
xset -dpms

[bonkersbobcat]

I have seen something similar to this. It happens very rarely, maybe once a month. When it happens SEQ will likely have been running for a couple of weeks.

What happens to me is that SEQ will hang and X (or the window manager) will loose the mouse. The desktop gets sort of "stuck". When in this state the mouse pointer is stuck and does not track movement of the mouse. I can use the alt-tab key sequence to switch windows, but only the window selector comes up - the on-top window does not change. I can use a ctrl-alt-backspace to exit the window manager and log back in, so the kernel is happy.

I haven't bothered to post it here, because in theory, nothing that SEQ does should be able to hang the underlying window manger. Normally this behavior would suggest looking into a problem with the window manager, but perhaps SEQ is pushing some button that it shouldn't be.

I am running a stock RedHat 9 installation on a PIII 500 with 512MB, using the default Gnome desktop. I think the problems first appeared with the RH9 installation.

I don't think I have seen it in the last two SEQ builds, but I haven't been playing as much lately so not sure if the problem has gone away or is just very intermittent.

[Amadeus]

I am running a stock RedHat 9 installation on a PIII 500 with 512MB, using the default Gnome desktop. I think the problems first appeared with the RH9 installation.


I think you're right..I don't think I noticed this before I upgraded to RedHat 9.0 as well.

[perlmonkey]

I've noticed problems like this too with RH9, but no longer, now that I run XD2 from Ximian. I think this was a problem with the screensaver trying to grab the X server for a 3D (e.g. Mesa) mode screensaver and failing. Usually, my machine would lock up UI-wise, but I could still log into the box remotely and all was well.

[Cryonic]

hmm, wonder if those systems are running ATI cards. We have issues here with ATI 7000 and 7500 cards and OpenGL graphics when hardware acceleration is turned on. This is one reason why 3D animation companies (like Newtek) like Nvidia still, fewer issues with OpenGL compliance.

[bonkersbobcat]

In my case the machine is a Dell machine with video down on the motherboard. It is a Intel 810 based (82810E) video chipset.

[perlmonkey]

Originally posted by Cryonic
hmm, wonder if those systems are running ATI cards. We have issues here with ATI 7000 and 7500 cards and OpenGL graphics when hardware acceleration is turned on. This is one reason why 3D animation companies (like Newtek) like Nvidia still, fewer issues with OpenGL compliance.

Yes, in my case it was ATI.