Opensuse won't start in normal mode after live update

Hello,

I’ve been using OpenSUSE 12.3 without big troubles since a while until now. Yesterday I decided that I will update the system to 13.1, using live update (in hindsight I should have sticked to the offline method… but it happened).
Anyway, since the update the system won’t boot completely. It reaches the default dark screen with the chameleon, but the login screen won’t come up if I boot on the regular way. It simply stops on that screen - the numlock is working, but I can’t even bring up a console with alt-Fx.
Meanwhile it boots to desktop (i use KDE) with the recovery option without a problem - I am writing this post from recovery also.

Being a linux noob, I am not really sure where should I start. I tried to make a new user, removing the nvidia vga drivers (I have found some posts that claimed it might cause such problems), reinstalling kde packages, but unfortunately no luck so far.

Here are the logs of a regular boot from /var/log/messages:

[messages - Pastebin.com](http://pastebin.com/Kbs1HPK1)

It simply stops after the 1149th line, even the hdd stops (i tried to leave it for more than 10 minutes, result is the same):

2013-12-25T09:09:14.735642+01:00 linux-daot systemd[1]: Started Update UTMP about System Runlevel Changes
2013-12-25T09:09:14.736349+01:00 linux-daot systemd[1]: Startup finished in 6.321s (kernel) + 33.121s (userspace) = 39.443s.

All suggestions would be appreciated, and if any other info is needed, ask away, will try to get it. Thanks in advance.

In my opinion, it’easier and quicker to do a fresh installation, keeping the home partition …

Well, I was thinking about something similar also… but then I get another problem: when I start the installation, the installer wants to use my entire hdd (it’s dual boot with windows - i use it about once in every 2 months, but it would hurt if i would lose it).

Warning: Your system states that it requires an EFI boot setup. Since the selected disk does not contain a GPT disk label, YaST will create a GPT label on this disk. You need to mark all partitions on this disk for removal. 

But I would go with whichever is easier - i’m okay with the re-installation and with fixing the current system aswell. I’m just a bit stuck both ways.

If you want to try to repair the system, you could dig around in /var/log/ and see what kinds of errors are being logged.

I would look at recent log files such as /var/log/Xorg.0.log (current boot/session) /var/log/Xorg.0.log.old (previous boot/session), /var/log/messages, /var/log/boot.log.

Since you are able to boot into recovery mode, maybe your graphics driver is not correctly setup/configured - the contents of Xorg.0.log.old might contain clues if your previous boot failed.

Oh yeahhhhh, back on track… thanks man, helped a lot. I have found a segmentation fault in x log. I googled a bit for that, it was a problem with xorg.conf. Or to be more exact, the problem was that it did not exist on my system:O : it was called “xorg.config.install”. I renamed it first for fun only to xorg.config - and suddenly it booted up normally :slight_smile:

Thanks again.

PS: Is there a way to mark threads as ‘solved’ here?

No, you should not need an xorg.conf.
And xorg.conf.install is the configuration for failsafe/recovery mode, you should not rename that, otherwise recovery mode won’t work anymore.

Your X segfault most likely is caused by the gfx driver though. Maybe you didn’t uninstall the nvidia driver correctly? Maybe nouveau doesn’t work on your system?
Could you post that Xorg.0.log(.old) with the segfault?
What gfx card do you have?

I have reverted the change I made about xorg.conf - and it booted once again correctly, I am unable to get the logs again. Meanwhile most probably you are right about the drivers, maybe they were not removed completely/correctly previously. I use an ASUS K55V series laptop with GeF 610M + Intel HD something (Optimus). Once I managed to get into the system (about 2 hours ago), I have uninstalled bumblebee (my previous tries with the drivers broke it also), and then I re-installed it completely along with the drivers it downloaded - it seems that solved the problem.

Thanks to you too, I guess you saved me from a heart attack from the future when I will want to use recovery again.

On 2013-12-25 10:06, skandigraun wrote:
>
> PiElle;2611281 Wrote:
>> In my opinion, it’easier and quicker to do a fresh installation, keeping
>> the home partition …
> Well, I was thinking about something similar also… but then I get
> another problem: when I start the installation, the installer wants to
> use my entire hdd (it’s dual boot with windows - i use it about once in
> every 2 months, but it would hurt if i would lose it).

You normally tell the installer to dump all its ideas, and just read the previous fstab file and use
those definitions.

But the next error baffles me:

> Code:
> --------------------
> Warning: Your system states that it requires an EFI boot setup. Since the selected disk does not contain a GPT disk label, YaST will create a GPT label on this disk. You need to mark all partitions on this disk for removal.
> --------------------

Humbug! It is working currently as traditional partition, so you do not need EFI.

>
> But I would go with whichever is easier - i’m okay with the
> re-installation and with fixing the current system aswell. I’m just a
> bit stuck both ways.

A possible way is to still do an offline upgrade on it, with the 13.1 DVD. It would erase completely
the nvidia drivers, too, and many other packages.


Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” (Elessar))

On 2013-12-25 10:36, skandigraun wrote:

> Oh yeahhhhh, back on track… thanks man, helped a lot. I have found a
> segmentation fault in x log. I googled a bit for that, it was a problem
> with xorg.conf. Or to be more exact, the problem was that it did not
> exist on my system:O : it was called “xorg.config.install”. I renamed it
> first for fun only to xorg.config - and suddenly it booted up normally
> :slight_smile:

That file is the same configuration as is used for the failsafe boot. Instead of renaming, just copy
it, and then you can boot normally - because the failsafe boot also disables all cpus cores but one
and is pretty slow.


Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” (Elessar))