[SOLVED (kind of]Add-repository->things have gone bad

[cogitans]

I just tried to upgrade my playonlinux through the repository. So I added the repository through the terminal which resulted in a key was missing. So I added the key, too. But the "apt-get upgrade" could not find anything (my previous playonlinux is 4.1.9 and the working version is something like 4.9). Furhtermore the Mint Updatemanager began to show errors. So I decided to remove the playonlinux-repository again.

After I removed the playonlinux-repository the system is kind of unstable. First symptom was that the sound was missing from the system. I did some attempts to fix the problem but nothing seemed to work. So I thought that a reboot might fix it. After a reboot I couldn't even get to the login-screen - I was stuck in the terminal which was saying "Checking ....".

I pressed ALT+F1 and logged in as the user and typed "startx". Now I'm back in the GUI. But the sound is still missing and I guess that a reboot will result in the same problems about the login-screen.

Now my question is: how do I fix these issues? And can it be a correct observation that the entire system is broken after a simple add-repository concerning playonlinux - and removal of it again?

(by the way I tried this, too:
sudo apt-get clean
sudo rm -r /var/lib/apt/lists/*
sudo touch /var/lib/apt/lists/lock
sudo mkdir /var/lib/apt/lists/partial
sudo apt-get clean
sudo apt-get update
)

Thanks in advance!

[cogitans]

Ah, I figured things out now.

I searched for "MDM" in Synaptic and things turned out that it wasn't installed. Then I installed it and did a reboot. Now the login-screen came automatically and after a login the sound is working again

However - it's beyond me how the MDM could get uninstalled from my attempt to add the PlayOnLinux-repository....(?)

[bigj231]

That is odd, but not the first time I've seen it happen. Glad you got it working. Could you edit the title of you original post to include [SOLVED] to let others know you got it fixed?

[cogitans]

Just to make things worse:
yesterday I did a modification of a mediaclip through Handbrake. During conversion the screen got about 20-30% more gamma. When the conversion had completed I though a reboot would fix the gamma-issue. So I did a reboot.

After the reboot the computer would yet AGAIN not go into login-mode. It went to terminal and kept printing some errors in a line at a time about the usb (I think it was). These errors was printed in a new line with about 3-5 seconds apart.

So I tried to login through another shell - ALT+F?. In some of these shells the error was printed out, too. In other shells I was able to begin debugging. So - the error just before the system went into terminal was that the screen couldn't be found. That was the same reason that I couldn't just type "startx" as X wasn't loaded. So I began to stop mdm and all that I learned from the previous case. But that didn't help. I even tried to uninstall mdm and then to install it again. But that didn't do any trick.

So I had to pull out my laptop and begin a search for a solution. Most of the solutions told me to go into the Xorg.conf and modify the driver. To change "nvidia" to "vs" or "ALSA" or something. I tried all the solutions but that's didn't do the trick either.

As time passed I was pretty much ready to do a complete reinstall of the OS (I've got an image of the system working - I learned that last time I had problems). But then finally I found a post online that I had to install the "Nvidia-current" through "apt-get install". So this was my last attempt to solve these issues.

The system did a lot of installing and then I did a reboot. And voila - I was presented by the loginscreen yet again. And now everything is working again.

Buttomline is that somehow the installation of the Nvidia-driver from Nvidia was smashed. It emerged when I was using Handbrake so I blame Handbrake (...?).
Then I had to install the Nvidiadriver from Ubuntu on top of it and after a reboot everything is okay again....strange!

By the way:
when I was troubleshooting the nvidia-driver-problem initially (the time before this one) the system told me that I couldn't install "NVIDIA-Linux-x86-304.64.run" because it was the 64bit-version of the driver and the kernel wouldn't accept it (my system is 32bit). So I downloaded "NVIDIA-Linux-x86-304.51.run" and succeeded.
Anyway, now that I've installed "nvidia-current" on top of everything and I go into "Administration->Nvidia X Server Settings" and I choose "NVIDIA X Server settings" I see that "X Server Information->NVIDIA Driver version" = 304.64.

This OS is getting more and more strange!!!

[bigj231]

Modifying your xorg.conf only helps if you can't start X. It sounds like your computer didn't even finish booting if it gave you one error at a time and wouldn't let you login in the virtual terminals. Sometimes the proprietary drivers have bugs. If handbrake was using your GPU (I thought it used the CPU to transcode?) it could have caused an issue, but that should have gone away on the next boot. Were you running it with superuser permissions?
Also, the driver naming scheme would suggest that the other file you downloaded should have been a 32 bit package. Never try to install 64 bit drivers on a 32 bit system (and vice-versa). The nvidia-current package is just the latest driver from nvidia, pre-packed and ready for use (hopefully). It will now upgrade if you use apt-get to upgrade your system. If you use the mint updater, you will need to check and make sure it doesn't update later (unless you are having problems of course). GPU drivers were my #1 headache on AMD hardware.

[cogitans]

Modifying your xorg.conf only helps if you can't start X.

Well, I kinda wrote that I couldn't start X when I got to try. I couldn't start X even when I reinstalled MDM.
That's why I tried to modify xorg.conf - which I was told through various post online.

It sounds like your computer didn't even finish booting if it gave you one error at a time and wouldn't let you login in the virtual terminals.

It DID finish booting. I got to the same point in the bootingorder as last time I had trouble loging in. The only difference this time was that the terminal kept on writing some number (maybe the hardware-address - I don't remember the exact term, but I was told to write a number in from of the graphiccard in the xorg.conf - mine was "0:0:2") followed by "USB can't be initiated" or something like that.
This line of error was being printed out over and over on new lines - just with 3-5 seconds apart. When I jumped to other terminal the error was printed out there, too. In some of the terminals it didn't though (in one case of booting I tried 6 different terminals before I found one that didn't write the line of error).

If handbrake was using your GPU (I thought it used the CPU to transcode?) it could have caused an issue, but that should have gone away on the next boot.

I think you're correct here. I didn't think about the fact that when using Handbrake then the GPU is being used and hence the graphicdriver. Maybe there was a bug in the driver - but I did think about that the driver was from Nvidia so they should be aware of all implementation. But then again - it's from Nvidia, not from Ubuntu, Mint or some other folks knowing about the Linux-implementation.

Were you running it with superuser permissions?

I assume it's still Handbrake we're talking about here. No, when launching Handbrake I do it from the default menus.

Also, the driver naming scheme would suggest that the other file you downloaded should have been a 32 bit package. Never try to install 64 bit drivers on a 32 bit system (and vice-versa).

Of course people are to use the correct driver concerning the bit to their system. I wouldn't even try installing 64bit-driver if I knew it was 64 bit. But I didn't see anything about "64 bit" when downloading the driver and I didn't think that the filename would consist of the bitnumber - I thought it would be the versionnumber.
But luckily the kernel was able to detect that it was a 64bit driver before it was installed. That's why I didn't install it either (I don't know it but I would NOT think that people even was able to install a 64bit driver on a 32bit system).

The nvidia-current package is just the latest driver from nvidia, pre-packed and ready for use (hopefully).

Oh, relly? I thought that the only way to obtain anything from Nvidia was through their website and do a compilation/installation yourself. But if "nvidia-current" is from Nvidia then that part is explained. But then again - when I had this issue initially I had a lot of problems installing a graphicdriver. My solution was to download a driver from Nvidia and install it separetly from debugmode. But aparantly the issue wasn't even about the driver it self (both drivers was from nvidia) - it must have been some system/kernel-issue and I was only able to install it from debugmode - not even from debugmode with webaccess so I could use apt-get to install nvidia-current.

GPU drivers were my #1 headache on AMD hardware.

I'm not using AMD anymore - this is Intel

[cogitans]

When I did a bootup this morning the screen was back to 20-30% extra gamma - the screen was too bright to watch.

So I did an apt-get remove mdm & apt-get remove nvidia-current & apt-get remove nvidia-current-updates

Then I did a reboot. When the computer had restarted and I logged in everything was viewable again. But then I opened my Liferea to view the news. When I clicked the first item in Liferea the brightness of the screen suddenly changed to bright again (+20-30 % gamma). As I'm fed up with troubleshooting this crazy issue I chose to open nvidia-settings and simply change the colour of the screen so that the content of the desktop is viewable again.

Ah, it's nice with a stable system that doesn't change it's settings when it's shut down over the night. I wish that Linux Mint 14 was one of those....

[cogitans]

Yaiiii - I fixed it

Today I lauched Synaptic and did a search for "nvidia". Then I did a complete removal of everything concerning "nvidia". I was hoping for the best surprice and did a reboot.

After the reboot the loginscreen came and I tried to login. But in every case I got the error "cinnamon can not be initiated" (or something like that concerning cinnamon). This probalby happened because the system had no drivers to show 3D. It would be fair to assume that the OS automatically would install some default driver in all circumstanses. I couldn't even get a terminal so I could reinstall the nvidia-stuff.

So I did a reboot into safemode. In here I upgraded the kernel to the newest generic versions (because when I installed the driver from Nvidia manually I checked the option to install something into the kernel so it would automatically update something on new versions. I don't think that the OS is that advanced that it unchecked this option when I uninstalled the driver - which is why I wanted to completelly overwrite the kernel to get rid of those changes) and the I did an "apt-get install nvidia-current". After a reboot...with fingers crossed...TADA: the screen and the resolution is back to normal.

Only "thing" is that in "Nvidia X server settings"->"X server information"->"Nvidia driver version"=304.64
I'm affraid that another problem is waiting in the future....

But for today I'm happy. I fixed this rather crazy issue. Although the troubleshooting was a so-and-so-solution then that's apparantly what is needed when the system isn't that stable after all.

[cogitans]

UPDATE:

And yet again this everlasting quest for a stable system concerning the graphics has evolved.

When I booted the system this morning the UpdateManager told me that there was a new driver - 304.84. "At last" was my initial thought. But as I had some work to do first I didn't update the system at first.

About 1 hour later the screen began to flicker. After a few flicks in 3-4 seconds the screen was yet AGAIN +20-30% gamma. So I did the driverupdate. And then a reboot.

But the gamma wasn't changed. So I repeated the process I discovered yesterday. And then I did a reboot.

But the gamma wasn't changed. So I had to do some new troubleshooting. And several reboots.

Finally I complettely removed everything concerning nvidia. Then I booted into safemode and installed "nvidia-173" and then a reboot. THEN the screen was back to normal.

Lets see how long the system is "stable" this time...

[cogitans]

Aaaand....we're back in the business of troubleshooting.

This morning the screen went back to being 20-30% too bright. So I updated to the nvidia-current-updates -> no change. Then I updated to the driver downloaded manually from Nvidia -> no change.

Hmm...maybe I'm forced to go buy a new computer although all the troubles with Linux Mint actually was an attempt to avoid having to spend money on a new computer