Graphical session randomly dies

I am not exactly sure what to call my problem. On my desktop I usually have two graphical sessions running, and switch back and forth with ctrl-alt-f7 & f8. From time to time, when I go to switch sessions, I will find myself at a login prompt. Sometimes it is the session at f7 that dies, other times f8. I have not been able to determine if a single application is causing the issue, as a couple times the session I was going to had no applications other than the DM running.

The error is very inconsistent also in that I can go days without an issue, then it will “decide” to have a bad day and crash every other time I switch back and forth, then “behave” again for a few days without restarting the system.

I am running Leap 42.1, Unity 5 on both sessions and using KDM. The system has an nVidia 8800gt and is running driver version 340.98 downloaded from nVidia.

What does the xorg log say for the session that has died?

Looking at Xorg.0.log.old

[237955.875] (EE)
[237955.875] (EE) Backtrace:
[237955.876] (EE) 0: /usr/bin/X (xorg_backtrace+0x48) [0x58b268]
[237955.876] (EE) 1: /usr/bin/X (0x400000+0x18f609) [0x58f609]
[237955.876] (EE) 2: /lib64/ (0x7f31ae3ae000+0x35140) [0x7f31ae3e3140]
[237955.876] (EE) 3: /usr/lib64/xorg/modules/drivers/ (0x7f31a81b3000+0x8a8bd) [0x7f31a823d8bd]
[237955.876] (EE) 4: /usr/lib64/xorg/modules/drivers/ (0x7f31a81b3000+0x5e1f39) [0x7f31a8794f39]
[237955.876] (EE) 5: /usr/lib64/xorg/modules/drivers/ (0x7f31a81b3000+0x5e6694) [0x7f31a8799694]
[237955.876] (EE) 6: /usr/lib64/xorg/modules/drivers/ (0x7f31a81b3000+0x5f6182) [0x7f31a87a9182]
[237955.876] (EE) 7: /usr/lib64/xorg/modules/drivers/ (0x7f31a81b3000+0x5edb61) [0x7f31a87a0b61]
[237955.876] (EE) 8: /usr/bin/X (xf86VTEnter+0x10b) [0x47b12b]
[237955.876] (EE) 9: /usr/bin/X (WakeupHandler+0x6d) [0x441add]
[237955.876] (EE) 10: /usr/bin/X (WaitForSomething+0x1d6) [0x588196]
[237955.876] (EE) 11: /usr/bin/X (0x400000+0x3cf61) [0x43cf61]
[237955.876] (EE) 12: /usr/bin/X (0x400000+0x40feb) [0x440feb]
[237955.876] (EE) 13: /lib64/ (__libc_start_main+0xf5) [0x7f31ae3cfb25]
[237955.876] (EE) 14: /usr/bin/X (0x400000+0x2c60e) [0x42c60e]
[237955.876] (EE)
[237955.876] (EE) Segmentation fault at address 0x30
[237955.876] (EE)
Fatal server error:
[237955.876] (EE) Caught signal 11 (Segmentation fault). Server aborting
[237955.876] (EE)
[237955.876] (EE)
Please consult the The X.Org Foundation support
for help.
[237955.876] (EE) Please also check the log file at “/var/log/Xorg.0.log” for additional information.
[237955.876] (EE)

Looks like something between the nvidia xorg driver and the X server isn’t working too well and generates a seg fault. If you haven’t already, you should address this with the folks at Nvidia.

Yeah, after going over the system with a friend that is better versed than me at diagnosing issues like this, all signs (well most of them at least) point at the nVidia drivers. Since I was not having issues when the system was still running 13.2 and an older version of the drivers I am going to roll them back and see if things are stable. If I still have issues, I have an AMD 5770 I am going to try in this machine.

Currently, it has been going strong for the last 4 days or so after having a… “rough day” where it decided to drop one of the sessions almost every other time I switched.

As to installing older drivers, do you (or anyone for that matter) think that using the drivers provided in the repo would make any difference compared to my usual method of downloading them right from nVidia?

Nothing wrong on installing from NVIDA but you do have to reinstall after any kernel or X stack update. If you use the NVIDIA repo then it will happen automagically

Reinstall the nvidia drivers, for your running kernel.