X started freezing
 

This week for the first time my Slackware64 14.0 system started freezing once every day/2 days. It always happens when I click on a link to a website from within Thunderbird. These are not dodgy websites - today it was docs.slackware.com, yesterday LQ. So I click on an url and things freeze. I can move a mouse pointer but the keyboard does not respond to Ctrl+Alt+Delete/Bksp (or any other window-manager specific shortcuts.) Yesterday I managed to close the system with REISUB.

Setup:
1. Nvidia binary blob - I haven't touched/upgraded it in at least a month
2. i3 window manager - hasn't been upgraded in a few months
3. Last week installed multilib - It suspiciously coincides with the time the freezes started appearing - it might just be a coincidence, though.

5. The system doesn't freeze as the music still plays in the background (mpd/ncmpcpp)
6. I can log in to that system with ssh
7. As I mentioned before, I can move the mouse.
8. Now I'm logged to that system via ssh. I'm trying to kill X remotely. I've tried init 3, no visible effect on the screen. I always startx. I successfully killed Firefox/Thunderbird - remotely ps aux shows they are killed. Yet I can still see them on a screen.
9. /tmp is mounted on tmpfs but it doesn't seem like I've run out of RAM (no swap)
10. I think when it happens I have the following programs open: firefox, thunderbird, spacefm, gimp, terminals
11. flash problem - I haven't upgraded flash in a good few months.

Tomorrow I'm not going to power on Thunderbird just to see if it happens again. If it doesn't help, I'll revert to a pure 64-bit system.

In the meantime, what else can I do to get to the bottom of it and/or kill X without shutting down the computer?

Thank you.

EDIT: I managed to kill X with killall -9 X and startx remotely. Well, that's not a solution to the problem but at least I could shut it down properly before going to sleep.

X started freezing
 

I seem to have similar problems, although I lose both mouse and keyboard and sometimes I can't connect with SSH.
No logs anywhere of course.
Doesn't happen on kernel-3.9 but does on both 3.10.7 and 3.11rc.
Have NVIDIA blob 325.15. Have down graded xorg-server from 1.4x to 1.3x.
Same as Sycamorex the freezes seem to happen in a browser( Seamonkey, Firefox or Konqueror) so disabled mozilla plugins and upgraded icedtea and flashplayer and machine ran faultlessly last night after that; but this morning wallop another freeze. WEnt out for half an hour and it was locked up when I got back. Could it be something to do with the screen being put into power save mode I wonder?


I am using KDE-4.10.5 as as far as I am aware there is a problem with 4.11 and the NVIDIA blob.

Mike

If its freezing - have you tried to turn up the heat ?? (sorry, couldn't help myself!)

Are you by any chance using a kvm-switch?
I have had something similar happening to me several times (_and_ I use a kvm-switch). If I reset the switch - I'm ok again.
I am running 14.0 with a 3.7.1-kernel (which otherwise seems to be extremely stable) and xfce.

sounds like teh video driver or the video hardware has crashed. I had such situations before, but I can't remember the context or any details. Re-starting X often does not help. The most aggressive re-initialization is done by "rmmod nvidia", then lsmod to check it has gone, followed by modprobe or startx. It sometimes even works. ;)

No, I'm not using kvm-switch. Thanks for your suggestion.

Thanks for the suggestion. I'll try it next time if nothing else works.

Another test you could try is copying the link and pasting it into your browser to open it, then trying to open it by clicking on it in Thunderbird.

It's a little laborious, but it would help determine whether the issue's related to clicking a link in Thunderbird or to opening the link in the browser.

Try CUDA GPU memtest to try to exclude a hardware issue:
http://sourceforge.net/projects/cudagpumemtest/

I've had similar symptoms and eventually I just got rid of the card, and suspect that it was faulty. I have tried both nvidia and nouveau drivers of all possible versions and there was no improvement. I have also tried various Xorg fixes.

I hope you find a better solution.

Does dmesg show anything strange?

also I would try to modify 3d and other acceleration settings. and look at xorg error log

recompiling the kernel with all Power Management and ACPI turned off seems to have fixed it for me ( although I'm sure there is a more correct way).2 days no freezes.


Mike

That is an extreme solution. Can you even shutdown properly ?

I realise it's extreme but I've run out of other things to try and can't begin to guess why kernel 3.9 is OK but 10 and 11 were unusable ( using the same .config)

You're right. The machine halts but doesn't power-off. I can live with that.


Mike

I haven't updated the thread in a couple of days as everything has been ok. Mind you my usage of that machine has been somewhat limited. I'm going to use it much more extensively over the next two days.

I have slack 14 and a Dell Inspiron 14 notebook, a Ubuntu one. I got more then 60% for my slack, but it freezes, the same way you are telling. I just stoped using it, now I am always using ubuntu. I will try the slack 14.1 when it is offered, and if things saty the same, I will probably install a Debian. I can stay with ubuntu more time or vI will get crazy and more dump.

I am currently investigating a possible kernel bug that may be causing this that was likely introduced with the 3.x kernels, because 2.6.x kernels seem to work ok.

2.9.x?

Sorry, 2.6.x.

You might also try to kill x with AltGr-SysRQ-r,AltGr-SysRQ-k.

Would it be an idea to update to current to check if that solves the freeze problem?

Ok. If it's 2.6.x instead of 3.9.x then I might have problems anyway.
Just upgraded from 3.7 to 3.10 this week :-P

Just an update - not that I complain but I don't experience any freezes any more. I have not done anything to fix it. It just stopped freezing. The only difference is that when the freezes happened I worked a lot with gimp and inkscape. I don't anymore. Perhaps it was related to that.

Nevermind, it wasn't a bug in the kernel, it was just bad hardware.