P.S. I also cannot figure out how to boot single user. Normally, I would press something like “select boot options” enter “1” and press “enter”, but that course of action is not available. If I try “sudo init 1” it drops down to single user, but without having rebooted. So, I’m where I was before anyway. I need to be able to boot up into single user and I cannot figure out how to do it. Thanks again.
Thank you. I know that. Two little problems: 1) I must be ready to pounce on that down arrow during that 8 second window (yes, I know I can edit /etc/default/grub and increase it, but still:) and 2) the original kernel hangs now, too.
You can switch back you going to Yast Software Management.
Search for “kernel”. Click the “Versions” button. And then click on the one you want to uninstall. I think the first click will give a green mark for reinstall, and a second click will change that to a red X for uninstall.
But you might want to first report as a bug. And perhaps it’s useful to keep the kernel around for answering questions raised about the bug report.
That is a possibility. Long ago, before I knew better, I picked the highly disparaged nVidia card that is not supported for “xinerama” by the open source driver. (It works, but it must be reconfigured after every restart. Another story.) So I use nVidia’s driver even though it “taints” the kernel.
Anyway, when I say, “hung” it means it gets past the console text and into an openSuSE 12.3 splash screen. As soon as it looks like it is taking too long, I can switch back to the console with Ctl-Alt-F1 and back to the splash with Ctl-Alt-F7, but it will soon freeze in whichever screen I happen to be in. The F1 screen is non-functional. It never gets to a login prompt. Since I don’t know what it thinks it is doing and since I cannot do anything, I call it “hung”. Full system hard reset required to get out of the state. (i.e. no magic key sequences work)
I have the same issue after the online update to Kernel 3.7.10-1.4. My computer boots to the black screen with the green leafy doodad on the right and STOPS!
No user log in screen. Nothing I’ve tried short of rolling it back works and that is not an option because I’ve read in this thread that does not work either, and so whats the point in doing the upgrade? It will boot in Kernel 3.7.10-1.4 failsafe. but nowhere else.
How does one re-install just the video driver? If you mean build and install the nVidia driver by running that self-extracting scripty thingy, yep. No change.
did you try booting to rescue mode?
Yep. I can use one of my two screens just fine.
did you try booting to a terminal ie run level 3?
I was able to take it down to run level 3 so that the scripty thingy would run. Yep. Did that under the rescue mode boot. I’ll need to try getting to multi-user mode/run level 3/init 3 now that I know the “type ‘e’ at the menu” grub2 magic. But unless I can get this working, I think I’m back to a re-install from the ground up.
Why don’t you just use the rpms from the nvidia repo? Those should still work even after a kernel update.
See SDB:NVIDIA drivers - openSUSE Wiki
If you install the driver “by running that self-extracting scripty thingy”, you have to recompile/reinstall the kernel module after every kernel update (with the new kernel running). See here: SDB:NVIDIA the hard way - openSUSE Wiki
If it actually worked when they first said it would, I would have. But it didn’t, so I haven’t been, so I didn’t. I could try it again.
If you install the driver “by running that self-extracting scripty thingy”, you have to recompile/reinstall the kernel module after every kernel update…]
Understood. Actually, it was the reason I was wanting the boot-into-multi-user information. That script only runs in run level 3. So I managed to coerce it into init 3 and rebuilt the driver. I am here because that didn’t work.
Well, I never had problems with those. (I have them installed on my girlfriend’s system, so they HAVE to work…;))
Understood. Actually, it was the reason I was wanting the boot-into-multi-user information. That script only runs in run level 3. So I managed to coerce it into init 3 and rebuilt the driver. I am here because that didn’t work.
Well, the reason should be found in your X logfile but I’m quite sure that the nvidia kernel module can’t be loaded.
Either boot to runlevel 3 after an unsuccessful boot and post /var/log/Xorg.0.log or boot into failsafe and post /var/log/Xorg.0.log.old .
You could also try to uninstall the driver completely and then install it again.
Or rename the file /etc/X11/xorg.conf if it exists, without it X should start and use the nouveau driver instead, if the nvidia driver fails to load.
Maybe it got fixed? I don’t often retry things unless someone says, “all better now”. I have too many bruises from retrying stuff in times past. Especially nouveau driver stuff.
Well, the reason should be found in your X logfile but I’m quite sure that the nvidia kernel module can’t be loaded.
There is this error entry:
[68.676] (EE) Failed to load /usr/lib64/xorg/modules/libglamoregl.so: /usr/lib64/xorg/modules/libglamoregl.so: undefined symbol: _glapi_tls_Context
That caused the “glamoregl” module to fail to load. I don’t get warm and fuzzies from it.
Nouveau won’t load 'cuz I poisoned it. As far as I know, you have to reconfigure nouveau on every X restart and I am waaaay to lazy to do that every day. I configured /etc/X11/… to work with nVidia and I surely do hope the kernel upgrade won’t invalidate the configuration.
Also, going back to the “old” log (I’m home now), the “vboxvideo” failed, as above, then this:
(EE) Screen 0 deleted because of no matching config section.
(EE) open /dev/dri/card0: No such file or directory
(EE) Screen 0 deleted because of no matching config section.
(EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not found)
[FONT=arial]You are right, but the nVidia driver wasn’t loaded after that. (This time I booted into single user and brought it to RL 5 by hand).
Xorg.o.log.old will be up for a week, if you want to plod through the whole thing: [/FONT]nvidia driver failed to load - Pastebin.com
[FONT=arial]For me, I make little out of it, other than the “(EE)” markings.[/FONT]
It fails to start GUI (probably because you need to re-install video drivers as I said) and there is long standing bug that it often fails to switch to tty1 after boot is finished and remains “stuck” at boot splash.
Press Alt-F1 or Ctrl-Alt-F1, you should see login prompt.
What’s nouveau got to do with the nvidia driver rpms?
As I said, they work fine here.
But personally I wouldn’t really trust the 1-click installer.
I would suggest:
Open YaST->Software Repositories, click on “Add”, select “Community Repositories” and activate the NVIDIA Repository.
Open YaST->Software Management, type nvidia into the search field and install those 3 packages: “nvidia-gfxG03-kmp-desktop” (or nvidia-gfxG03-kmp-default, if you’re using kernel-default), “xorg-x11-nvidiaG03” and “nvidia-computeG03” (should be selected automatically)
(I’m assuming you need the G03 driver, if not replace the “G03” accordingly)
This should just work, but be sure to remove your hand-installed driver before…
There is this error entry:
That caused the “glamoregl” module to fail to load. I don’t get warm and fuzzies from it.
Nouveau won’t load 'cuz I poisoned it. As far as I know, you have to reconfigure nouveau on every X restart and I am waaaay to lazy to do that every day. I configured /etc/X11/… to work with nVidia and I surely do hope the kernel upgrade won’t invalidate the configuration.
Also, going back to the “old” log (I’m home now), the “vboxvideo” failed, as above, then this:
(EE) Screen 0 deleted because of no matching config section.
(EE) open /dev/dri/card0: No such file or directory
(EE) Screen 0 deleted because of no matching config section.
(EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not found)
[FONT=arial]You are right, but the nVidia driver wasn’t loaded after that. (This time I booted into single user and brought it to RL 5 by hand).
Xorg.o.log.old will be up for a week, if you want to plod through the whole thing: [/FONT]nvidia driver failed to load - Pastebin.com
[FONT=arial]For me, I make little out of it, other than the “(EE)” markings.[/FONT]
That’s the log from the failsafe boot, that doesn’t help here.
Be sure to do this:
do a normal boot, X should fail.
then reboot into failsafe and post /var/log/Xorg.0.log.old
The failsafe boot tries to use a variety of drivers, so those error messages don’t really matter.
It’s normal that vboxvideo can’t load, because you’re not running inside VirtualBox but on real hardware.
But there seems to be a mismatch between glamor and your Mesa packages. But I don’t think that would cause the nvidia driver to fail though. It would just prevent you from using glamor, which is 2d acceleration through OpenGL. I’m not sure if that is even applicable to the nvidia driver…
Maybe that error msg is even caused by the nvidia driver, since it replaces some parts of Mesa.
In my experience with this problem, Ctl-Alt-F1 switched you to the console screen, but it never got as far as the login prompt before locking up (“hanging”). The only thing I was able to do was switch back and forth between the console and GUI “desktops” (probably wrong word) – until lockup was complete. At that point, the machine only responded to a reset or power cycle.
@wolfi323 – I’ll need to try that most likely Thursday. Tue/Wed are grueling days and going home to sit at a computer some more is not appealing. Thank you for all your suggestions and I will follow up in a couple of days. Regards, Bruce
Just a fear that someone would say, “Well, use the nouveau driver.”
I wouldn’t really trust the 1-click installer. I would suggest:
Open YaST->Software Repositories, click on “Add”, select “Community Repositories” and activate the NVIDIA Repository.
Hmm. For openSuSE 12.3 I only see “openSUSE BuildService” and a list that does not include nVidia. I’d need a URL to configure it.
This should just work, but be sure to remove your hand-installed driver before…
Glad I didn’t.
… the [pastebin] log [is] from the failsafe boot, that doesn’t help here.
Good news and bad news: Without knowing exactly what it is that I did, somehow the nVidia driver is loading now.
$ lsmod|fgrep -i nvidia
nvidia 9174778 34
Whatever it is you are thinking about asking me, the answer is, “I don’t know.” Still, thank you so much for pointing me off in ways that will help “next time”.
Thank you for all your suggestions and I will follow up in a couple of days. Regards, Bruce
Just a fear that someone would say, “Well, use the nouveau driver.”