broken apt + fail to boot w/ init: nvidia-persistenced
 

X won’t start. I can’t even get to gdm on boot. When I boot, it’s just a black screen. I used VTs to get my dmesg and to tinker with apt. My problem might have something to do with a recent proprietary nvidia driver update.

Here is my dmesg:
https://www.dropbox.com/s/8coqimlgqg...c2k14.txt?dl=0
Take note near the end where it says:
I Googled that line and came across some forum threads related to corrupted nvidia drivers. As per some of the advice for other forum participants, I tried purging my system of everything nvidia with sudo apt-get purge nvidia-*, however I got this:
So apparently there aren’t very many nvidia proprietary drivers installed. I then tried installing an nvidia driver. I picked 319 so as to avoid the buggy bleeding edge (like 33x version). Here is me trying to install 319:
So I do have nvidia-319 already installed? If that’s true, then why when I tried to purge, did it say that it isn’t installed. Back up there it does say:What is going on? As you can see, there are some corrupted EFL packages in my system, but I don’t think that has anything to do with x not being able to start. I entered sudo apt-get install -f anyways. See here:
To capture these shell messages, I used sudo apt-get -f install -y >>file.txt which for some reason broke half way through and didn’t finish capturing the rest of the output for my apt-get -f install. Oh well.

Anyways, is my broken package tool related to my experience with being unable to boot? What’s wrong with my system and what do I have to do to get x to start? Is there any other information I can provide to help you ppl help me?

Hi,

what is the state of your packages now? Did you run "apt-get -f install" again?

To debug X11 problems you should investigate the X11 log file, usually found at /var/log/Xorg.0.log You probably have a /etc/X11/xorg.conf that is instructing X11 to use drivers that don't exist on your system any more.

Evo2.

Thanks evo2. You said:

Here are the Xorg.x.log files in my /var/log/ directory: https://www.dropbox.com/s/v1w0jcmke1...14.tar.gz?dl=0
I hesitated to include them in my initial post because for some reason I see that they are all dated 2014-04-13. I figured that they wouldn't be helpful today in December 20014. Alas, here they are. What can you make of them?

I did run apt-get -f install. I tried to capture the output with 'sudo apt-get -f install -y >>file.txt' as I said. But the capture breaks mid way through.

Hi,

Xorg-troubleshootDec2k14.tar.gz does not appear to be a gzipped tar file. My system seems to think it is a rar file but unrar failled to extract its contents. You need to a log file that corresponds to when you last tried to start X11 (and after you removed all the nvidia packages).

I understand that your tried to redirect the output, but try without redirecting the output and read what it says.

Evo2.

Here is a compressed rar file with my Xorg.x.log contents: https://www.dropbox.com/s/yufrdvyd8w...c2k14.rar?dl=0
Sorry for the confusion with the tar ball initially.

I can boot into a live environment, chroot into my system and then paste the output of 'sudo apt-get -f install -y'. I'll get to that in a few days (maybe tomorrow if I am really lucky).

Thanks, Evo2.

Hi,
I don't know what you are using to crate your rar files (or why you are using rar files), but unrar on my Debian Jessie system is unable to unpack it.
Seems I've missed something here. The system doesn't boot? If that is the case we should completely forget about your X11 problems for now. Please clarify the situation.

Evo2.

My apologies for the confusion, Evo2. This post should clarify the two issues at hand.

Here is a proper tallball of all my Xorg.x.logs: https://www.dropbox.com/s/ar7v36i8io...VD.tar.gz?dl=0

And here is the output of sudo apt-get -f install directly from my chroot shell: https://paste.ee/p/VUQ9e

Hi,
For some reason all those log files have exactly the same timestamp so I'm not sure which is the most recent one. Looking in Xorg.0.log I see that X is trying to load some nvidia drivers that you have uninstalled. You might be able to get X to run using free drivers if you move your config file out of the way. For example
Ouch. You have conflicts between enlightenment packages. I notice also that you have a significant number of different sources. It looks like you have tried to install enlightenment from more than one place. I recommend uninstalling all the enlightenment packages and removing the bohdi sources. Or you can try to use "apt-cache policy" to work out where the packages are coming from and then pick and choose which ones to remove. Mixing different 3rd party sources is the cause of most apt dependency problems that I see in the wild.

For a quick fix (that may not work) you could try remove just the offending package
and while you're at it
But anyway, can you explain your situation in regard to doing this in a chroot instead of booting into the OS?

Evo2.

Booting directly into my O/S into shell without a GUI (b/c X is broke), I noticed that I didn't have an xorg.conf in my /etc/X11. All I had was an xorg.conf.failsafe file - - a file I have never seen or heard of before. Strange. So I moved that file by adding a .bak file extension. I attempted startx and it didn't work. I held my breath and rebooted but turned out that the problem persists. I still can't load an x session and gdm still won't launch.


Here is my apt-cache: https://paste.ee/p/3tnNw

Nah, that didn't work. See here: https://paste.ee/p/O1eGK

Yes, I will explain that when I am back on Windows 7.

I chrooted into my box from a live CD so that I could copy and paste the output from sudo apt-get install -f. Remember? I was trying to capture the output by redirecting to a text file but it stopped half way through. Initially when I tried to troubleshoot this problem, I had booted in my native Ubuntu installation. Since X was broken, I was redirecting my output to text files and them copying over to my Dropbox folder. I started Dropbox in my shell and when I booted into my Windows 7 O/S, there were my text files. This also explains why all my Xorg.0.log files all have the same time stamps. I had copied them over to my Dropbox folder and synced them to the cloud. When I arrived on Windows 7, the file system stamped them in the present day. In retrospect, I prolly shoulda zipped them first and then transfer the zip file. Oh well.

By the way, here are the contents of my /etc/X11: https://www.dropbox.com/s/jjlgpm77ys...ic.tar.gz?dl=0

Thank you Evo2 for helping me out.

Hi,

ok regarding the chroot approach I think if you use the "script" command to record your terminal sessions.

Regarding your X config I see xorg.conf.failsafe.bak dated Dec 14. It tries to use the nvidia driver. Presumably this was in place as xorg.conf when you were unable to start X and was the cause of your problems. Please try to start X again and post the new /var/log/Xorg.0.log (or similar) that is created.

Evo2.

I'll keep this in mind for next time. Thanks for tipping this feature my way.

See attached for my latest Xorg.0.log.
For those not members of this forum who still want access to the attached file, here is a dropbox link: https://www.dropbox.com/s/8gololwqmq...org.0.log?dl=0

There's no /etc/X11/xorg.conf file, by default, in any recent *buntu installation. They stopped providing it several years back. However, one can be created manually, owned by root with 644 permissions, and will be used if it exists.

Your most recent log file indicates a problem with an Nvidia module; it also indicates that the system is loading a noveau module also, which is the open-source driver for Nvidia cards. I'm not quite sure from where these are being called, but would look first at /etc/modprobe.d for conf files containing their names.

Meanwhile, you might find it helpful to get to the grub menu at boot time, by pressing the left shift key as soon as the initial BIOS screen goes away, then selecting the topmost entry, pressing "e" to enter editing mode, scrolling down to the line containing "quiet splash" and adding "nomodeset" as a third word to that string, finally pressing CTRL-X to continue booting. This MAY (or may not) let you get into the desktop to do some troubleshooting direct.

Cheers for the advice, jim

I tried appending the nomodeset parameter via grub (to the line with quiet splash) and X still wouldn’t start. Here is the Xorg.0.log produced after I tried startx in a VT shell: https://www.dropbox.com/s/wu9c8134k2...c2k14.log?dl=0

There was no modprobe.d file in /etc/ Rather, nomodprobe.d was a directory. Here were the contents: https://www.dropbox.com/s/v9qmcy5u7g....d.tar.gz?dl=0

The contents of nvidia-331_hybrid.conf.dpkg-new are as follows:

I have Ubuntu running perfectly now. The solution was simple.

I commented out my bodhi packages in my sources and then ran sudo apt-get autoremove -f .

This automatically installed the latest kernel and then built an nvidia kernel module. This was previously getting blocked by the broken efl packages.

Thanks, jim and evo2.