I was using opensuse 13.1 for quite a while and upgraded it to 13.2 about a month ago. It happened that the OS had some trouble in starting and said it failed to find the screen. I uninstalled some NVIDIA drivers (my graphic card is a NVIDIA GT640), now I can not log in any longer. Every time I keyed user name and password in kdm , I was then put back to the kdm interface. The error information was “Problem loading in-kernel X.509 certificate (-19)” .
What do you think I can do to rescue my system?
The message means you are trying to run unsigned modules, which may be disallowed in UEFI secure boot environements. Try to fix things with UEFI secure boot disabled. The message could also be unrelated to your graphics issue. Pasting your /var/log/Xorg.0.log may help approaching it.
What exactly do you mean with “I uninstalled some NVIDIA drivers”?
That error message is irrelevant, and unrelated to your problem. And you do not get it when trying to login, but rather at the beginning of the boot process. You just see it when KDM restarts, as there’s a temporary switch to text mode.
But it would suggest you are using kernel 3.16.7-13, an incompatible kernel update that got revoked.
In any way, your problem is probably caused by a not-working nvidia driver (maybe because of that incompatible kernel) . Or you did not fully uninstall it and nouveau is broken now as well. Or nouveau has problems with your card.
What do you think I can do to rescue my system?
Try booting to the previous kernel in “Advanced Options” in the boot menu, and/or try booting to “Recovery Mode” there.
If you still cannot login to KDE, choose IceWM on the login screen, that should give you a graphical system at least.
You might try to run KDE inside IceWM (just type “startkde” into an xterm window), maybe you get a meaningful error message.
Then I would advise you to uninstall Kernel 3.16.7-13 (use YaST’s “Versions” tab for that), and reinstall the nvidia driver packages if necessary.
I followed wolfi3232’ suggestion and logged in viw IceWM , but I was unable to open /var/log/Xorg.0.log with any of these softwares: gedit, Beaver or leaf, the OS gave me a message “fail to load libnvidia-glcore.so.304.76, No such file or directory”.
Finally I used vi and found such the same message “fail to load libnvidia-glcore.so.304.76, No such file or directory”, I think this might be the responsible for my problem
I followed your suggestion and logged in viw IceWM , but I was unable to open /var/log/Xorg.0.log with any of these softwares: gedit, Beaver or leaf, the OS gave me a message “fail to load libnvidia-glcore.so.304.76, No such file or directory”.
Finally I used vi and found such the same message “fail to load libnvidia-glcore.so.304.76, No such file or directory”, I think this might be the responsible for my problem
[/QUOTE]
Then I would advise you to uninstall Kernel 3.16.7-13 (use YaST’s “Versions” tab for that), and reinstall the nvidia driver packages if necessary.[/QUOTE]
Thank you very much for you help. But I can not find the “Versions” tab in YaST.
[QUOTE]
Then I would advise you to uninstall Kernel 3.16.7-13 (use YaST’s “Versions” tab for that), and reinstall the nvidia driver packages if necessary.
Thank you very much for you help. But I can not find the “Versions” tab in YaST.[/QUOTE]
For the version of YaST that comes with KDE (QT based I believe) this is what one sees wrt versions :
It’s definitely one aspect of your problem. The system tries to load the nvidia libraries but cannot find them because they don’t exist apparently.
This is not related to the kernel, but rather because you “uninstalled some NVIDIA drivers”.
Apparently you only uninstalled part of the driver, and have other parts still installed.
So, how did you install the driver in the first place? Via the packages from the nvidia repo? Or rather the .run installer from nvidia’s homepage?
And that error message refers to the G02 driver which does not even support your card. And 304.76 is even a stone-old version of that (the current one is 304.125), and won’t work at all on openSUSE 13.2.
[EDIT]Ok, sorry, G02 would actually support your card, it does support cards upto GeForce 600 series, but the rest still applies[/EDIT]
Or did the error message rather say 340.76? That would be the current G03 driver.
So, all in all I would recommend to uninstall all nvidia related packages too, in addition to the kernel 3.16.7-13.2.
Search in YaST for “nvidia” and remove all packages that have nvidia in the name (not xf86-video-nouveau e.g.!).
If you need help, please post the output of this, and I’ll give you a shell command to run:
rpm -qa | grep nvidia
(this only applies if you installed the driver via the RPM packages)
An alternative way to uninstall the kernel 3.16.7-13.2 would be this:
sudo zypper rm kernel-desktop-3.16.7-13.2
Afterwards, you can of course install the driver again, but make sure you install either G03 or (better) G04.
On second thought:
Considering that you likely had 340.76 installed actually (not 304.76) and probably are using the nvidia repo, I’d recommend to run this to fix your system:
If you get error messages, please post them.
If that doesn’t help, please answer the questions from my previous post.
And since you upgraded from 13.1 (that’s when your problems started, right?), also delete the directories /usr/lib/xorg/modules/update/ /usr/lib64/xorg/modules/update/ if they exist, they might contain left-overs of the older 13.1 driver.
Sorry, I forgot about something: you should reboot after you removed the kernel, before you install the nvidia driver.
Otherwise you might still run the 3.16.7-13.2 kernel, and the installation of the driver will fail because the corresponding devel files are not found for the running kernel… Completely removing the nvidia driver might be a good idea too to be on the safe side.
