Stuck at login + nvidia* purged

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP








up vote
0
down vote

favorite












This question is popular however I searched and couldn't find a solution that fits unless I missed it somewhere. Please do not downvote before reading.



I have nvidia-387 driver installed and my system was working. I wanted to install cuda so I went with nvidia-390. This led to the login loop problem which I did not try to solve with this solution.



So I moved back to the nvidia-387 but i was stuck at login. I tried nomodeset and nouveau.modeset=0 but none worked on the latest kernel I was using. nouveau.modeset=0 did work on the 2nd kernel. So I went on and purged nvidia*. Now I am faced with the same problem on both kernels.



I am left with one kernel with which I can normally login.



How can I fix the other kernel(s) preferably without messing with the working one so I could get back to installing cuda?



System:



  • Ubuntu Gnome 16.04

  • Dell XPS 15 9560 with

  • NVIDIA GTX 1050

Additional info



On the working kernel, running sudo ubuntu-drivers list I get:



nvidia-384
nvidia-387
nvidia-390


enter image description here



apt-cache policy nvidia-* (also on working kernel; external monitor not working) output: https://paste.ubuntu.com/p/VsQPVCZfK4/



Update 1



Kernels listed in grub menu




  • 4.14.0-38: the one I was using with nvidia-387 but couldn't turn on GPU from nvidia-seettings. It suffered from this problem. I also installed bumblebee and switched to nvidia-390 before rebooting the system and ending up with the the black screen following the startup splash screen (even with nomodeset)

  • 4.13.0-37


  • 4.3.0-119: I was able to login to this yesterday but not after I tried ubuntu-drivers autoinstall (how silly of me; I expected to be asked for a confirmation before installation) which also installed nvidia-390 and can no longer login.

Now I can no longer login to any of the above listed kernels. My system is encrypted. If there's a way to fix the above issue (maybe install the latest upstream kernel?) that would be great. If not how can I reinstall the system while preserving or backing up my home directory?



Update 2



I manged to get back in using the first kernel. To do this I used root in recovery mode to move /etc/modprobe.d/bumblebee.conf to /etc/modprobe.d/bumblebee.conf.bk



The system is using nvidia-390 however the nvidia-settings don't show a GPU option.







share|improve this question






















  • Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
    – David Foerster
    Apr 20 at 22:49










  • @DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
    – user10853
    Apr 21 at 0:00






  • 1




    You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
    – WinEunuuchs2Unix
    Apr 21 at 0:04










  • @WinEunuuchs2Unix I updated my question with the asked details.
    – user10853
    Apr 21 at 18:58














up vote
0
down vote

favorite












This question is popular however I searched and couldn't find a solution that fits unless I missed it somewhere. Please do not downvote before reading.



I have nvidia-387 driver installed and my system was working. I wanted to install cuda so I went with nvidia-390. This led to the login loop problem which I did not try to solve with this solution.



So I moved back to the nvidia-387 but i was stuck at login. I tried nomodeset and nouveau.modeset=0 but none worked on the latest kernel I was using. nouveau.modeset=0 did work on the 2nd kernel. So I went on and purged nvidia*. Now I am faced with the same problem on both kernels.



I am left with one kernel with which I can normally login.



How can I fix the other kernel(s) preferably without messing with the working one so I could get back to installing cuda?



System:



  • Ubuntu Gnome 16.04

  • Dell XPS 15 9560 with

  • NVIDIA GTX 1050

Additional info



On the working kernel, running sudo ubuntu-drivers list I get:



nvidia-384
nvidia-387
nvidia-390


enter image description here



apt-cache policy nvidia-* (also on working kernel; external monitor not working) output: https://paste.ubuntu.com/p/VsQPVCZfK4/



Update 1



Kernels listed in grub menu




  • 4.14.0-38: the one I was using with nvidia-387 but couldn't turn on GPU from nvidia-seettings. It suffered from this problem. I also installed bumblebee and switched to nvidia-390 before rebooting the system and ending up with the the black screen following the startup splash screen (even with nomodeset)

  • 4.13.0-37


  • 4.3.0-119: I was able to login to this yesterday but not after I tried ubuntu-drivers autoinstall (how silly of me; I expected to be asked for a confirmation before installation) which also installed nvidia-390 and can no longer login.

Now I can no longer login to any of the above listed kernels. My system is encrypted. If there's a way to fix the above issue (maybe install the latest upstream kernel?) that would be great. If not how can I reinstall the system while preserving or backing up my home directory?



Update 2



I manged to get back in using the first kernel. To do this I used root in recovery mode to move /etc/modprobe.d/bumblebee.conf to /etc/modprobe.d/bumblebee.conf.bk



The system is using nvidia-390 however the nvidia-settings don't show a GPU option.







share|improve this question






















  • Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
    – David Foerster
    Apr 20 at 22:49










  • @DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
    – user10853
    Apr 21 at 0:00






  • 1




    You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
    – WinEunuuchs2Unix
    Apr 21 at 0:04










  • @WinEunuuchs2Unix I updated my question with the asked details.
    – user10853
    Apr 21 at 18:58












up vote
0
down vote

favorite









up vote
0
down vote

favorite











This question is popular however I searched and couldn't find a solution that fits unless I missed it somewhere. Please do not downvote before reading.



I have nvidia-387 driver installed and my system was working. I wanted to install cuda so I went with nvidia-390. This led to the login loop problem which I did not try to solve with this solution.



So I moved back to the nvidia-387 but i was stuck at login. I tried nomodeset and nouveau.modeset=0 but none worked on the latest kernel I was using. nouveau.modeset=0 did work on the 2nd kernel. So I went on and purged nvidia*. Now I am faced with the same problem on both kernels.



I am left with one kernel with which I can normally login.



How can I fix the other kernel(s) preferably without messing with the working one so I could get back to installing cuda?



System:



  • Ubuntu Gnome 16.04

  • Dell XPS 15 9560 with

  • NVIDIA GTX 1050

Additional info



On the working kernel, running sudo ubuntu-drivers list I get:



nvidia-384
nvidia-387
nvidia-390


enter image description here



apt-cache policy nvidia-* (also on working kernel; external monitor not working) output: https://paste.ubuntu.com/p/VsQPVCZfK4/



Update 1



Kernels listed in grub menu




  • 4.14.0-38: the one I was using with nvidia-387 but couldn't turn on GPU from nvidia-seettings. It suffered from this problem. I also installed bumblebee and switched to nvidia-390 before rebooting the system and ending up with the the black screen following the startup splash screen (even with nomodeset)

  • 4.13.0-37


  • 4.3.0-119: I was able to login to this yesterday but not after I tried ubuntu-drivers autoinstall (how silly of me; I expected to be asked for a confirmation before installation) which also installed nvidia-390 and can no longer login.

Now I can no longer login to any of the above listed kernels. My system is encrypted. If there's a way to fix the above issue (maybe install the latest upstream kernel?) that would be great. If not how can I reinstall the system while preserving or backing up my home directory?



Update 2



I manged to get back in using the first kernel. To do this I used root in recovery mode to move /etc/modprobe.d/bumblebee.conf to /etc/modprobe.d/bumblebee.conf.bk



The system is using nvidia-390 however the nvidia-settings don't show a GPU option.







share|improve this question














This question is popular however I searched and couldn't find a solution that fits unless I missed it somewhere. Please do not downvote before reading.



I have nvidia-387 driver installed and my system was working. I wanted to install cuda so I went with nvidia-390. This led to the login loop problem which I did not try to solve with this solution.



So I moved back to the nvidia-387 but i was stuck at login. I tried nomodeset and nouveau.modeset=0 but none worked on the latest kernel I was using. nouveau.modeset=0 did work on the 2nd kernel. So I went on and purged nvidia*. Now I am faced with the same problem on both kernels.



I am left with one kernel with which I can normally login.



How can I fix the other kernel(s) preferably without messing with the working one so I could get back to installing cuda?



System:



  • Ubuntu Gnome 16.04

  • Dell XPS 15 9560 with

  • NVIDIA GTX 1050

Additional info



On the working kernel, running sudo ubuntu-drivers list I get:



nvidia-384
nvidia-387
nvidia-390


enter image description here



apt-cache policy nvidia-* (also on working kernel; external monitor not working) output: https://paste.ubuntu.com/p/VsQPVCZfK4/



Update 1



Kernels listed in grub menu




  • 4.14.0-38: the one I was using with nvidia-387 but couldn't turn on GPU from nvidia-seettings. It suffered from this problem. I also installed bumblebee and switched to nvidia-390 before rebooting the system and ending up with the the black screen following the startup splash screen (even with nomodeset)

  • 4.13.0-37


  • 4.3.0-119: I was able to login to this yesterday but not after I tried ubuntu-drivers autoinstall (how silly of me; I expected to be asked for a confirmation before installation) which also installed nvidia-390 and can no longer login.

Now I can no longer login to any of the above listed kernels. My system is encrypted. If there's a way to fix the above issue (maybe install the latest upstream kernel?) that would be great. If not how can I reinstall the system while preserving or backing up my home directory?



Update 2



I manged to get back in using the first kernel. To do this I used root in recovery mode to move /etc/modprobe.d/bumblebee.conf to /etc/modprobe.d/bumblebee.conf.bk



The system is using nvidia-390 however the nvidia-settings don't show a GPU option.









share|improve this question













share|improve this question




share|improve this question








edited Apr 22 at 13:24

























asked Apr 20 at 18:30









user10853

4711423




4711423











  • Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
    – David Foerster
    Apr 20 at 22:49










  • @DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
    – user10853
    Apr 21 at 0:00






  • 1




    You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
    – WinEunuuchs2Unix
    Apr 21 at 0:04










  • @WinEunuuchs2Unix I updated my question with the asked details.
    – user10853
    Apr 21 at 18:58
















  • Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
    – David Foerster
    Apr 20 at 22:49










  • @DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
    – user10853
    Apr 21 at 0:00






  • 1




    You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
    – WinEunuuchs2Unix
    Apr 21 at 0:04










  • @WinEunuuchs2Unix I updated my question with the asked details.
    – user10853
    Apr 21 at 18:58















Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
– David Foerster
Apr 20 at 22:49




Does CUDA not work with the 387 driver series for some reason? Could you please edit your question to include the output of apt-cache policy nvidia-*? Thanks.
– David Foerster
Apr 20 at 22:49












@DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
– user10853
Apr 21 at 0:00




@DavidFoerster Updated. I did not say that. I though it would be good to use the latest version. Apparently, it wasn't.
– user10853
Apr 21 at 0:00




1




1




You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
– WinEunuuchs2Unix
Apr 21 at 0:04




You don't actually list the kernel versions that don't work and the one that does work. I think there were problems with nVidia supporting Meltdown and Spectre changes made for newest kernels which other people might spot if the relevant version numbers are in your question. Also you don't mention your nVidia model number and there may be problems with specific models others might spot.
– WinEunuuchs2Unix
Apr 21 at 0:04












@WinEunuuchs2Unix I updated my question with the asked details.
– user10853
Apr 21 at 18:58




@WinEunuuchs2Unix I updated my question with the asked details.
– user10853
Apr 21 at 18:58










1 Answer
1






active

oldest

votes

















up vote
0
down vote













I ended up grabbing the latest version of Ubuntu Gnome and reinstalling (preserving the encrypted /home).



Additional Drivers showed me only one nvidia driver this time: nvidia (proprietary, tested) which I used. As far as I can see the GPU seems to be working as the NVIDIA settings shows more options I didn't have before and the Performance mode is now selected. I added screenshots below.



I tested switching between Performance and Power Saving modes and it works however I need to reboot the system counter to logging out and back in as the NVIDIA settings says.



Also, battery life is down drastically from 5-6 hours on Power Saving mode to 2.5 hours on Performance mode. This is expected.



enter image description hereenter image description here






share|improve this answer






















    Your Answer







    StackExchange.ready(function()
    var channelOptions =
    tags: "".split(" "),
    id: "89"
    ;
    initTagRenderer("".split(" "), "".split(" "), channelOptions);

    StackExchange.using("externalEditor", function()
    // Have to fire editor after snippets, if snippets enabled
    if (StackExchange.settings.snippets.snippetsEnabled)
    StackExchange.using("snippets", function()
    createEditor();
    );

    else
    createEditor();

    );

    function createEditor()
    StackExchange.prepareEditor(
    heartbeatType: 'answer',
    convertImagesToLinks: true,
    noModals: false,
    showLowRepImageUploadWarning: true,
    reputationToPostImages: 10,
    bindNavPrevention: true,
    postfix: "",
    onDemand: true,
    discardSelector: ".discard-answer"
    ,immediatelyShowMarkdownHelp:true
    );



    );













     

    draft saved


    draft discarded


















    StackExchange.ready(
    function ()
    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2faskubuntu.com%2fquestions%2f1026762%2fstuck-at-login-nvidia-purged%23new-answer', 'question_page');

    );

    Post as a guest






























    1 Answer
    1






    active

    oldest

    votes








    1 Answer
    1






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes








    up vote
    0
    down vote













    I ended up grabbing the latest version of Ubuntu Gnome and reinstalling (preserving the encrypted /home).



    Additional Drivers showed me only one nvidia driver this time: nvidia (proprietary, tested) which I used. As far as I can see the GPU seems to be working as the NVIDIA settings shows more options I didn't have before and the Performance mode is now selected. I added screenshots below.



    I tested switching between Performance and Power Saving modes and it works however I need to reboot the system counter to logging out and back in as the NVIDIA settings says.



    Also, battery life is down drastically from 5-6 hours on Power Saving mode to 2.5 hours on Performance mode. This is expected.



    enter image description hereenter image description here






    share|improve this answer


























      up vote
      0
      down vote













      I ended up grabbing the latest version of Ubuntu Gnome and reinstalling (preserving the encrypted /home).



      Additional Drivers showed me only one nvidia driver this time: nvidia (proprietary, tested) which I used. As far as I can see the GPU seems to be working as the NVIDIA settings shows more options I didn't have before and the Performance mode is now selected. I added screenshots below.



      I tested switching between Performance and Power Saving modes and it works however I need to reboot the system counter to logging out and back in as the NVIDIA settings says.



      Also, battery life is down drastically from 5-6 hours on Power Saving mode to 2.5 hours on Performance mode. This is expected.



      enter image description hereenter image description here






      share|improve this answer
























        up vote
        0
        down vote










        up vote
        0
        down vote









        I ended up grabbing the latest version of Ubuntu Gnome and reinstalling (preserving the encrypted /home).



        Additional Drivers showed me only one nvidia driver this time: nvidia (proprietary, tested) which I used. As far as I can see the GPU seems to be working as the NVIDIA settings shows more options I didn't have before and the Performance mode is now selected. I added screenshots below.



        I tested switching between Performance and Power Saving modes and it works however I need to reboot the system counter to logging out and back in as the NVIDIA settings says.



        Also, battery life is down drastically from 5-6 hours on Power Saving mode to 2.5 hours on Performance mode. This is expected.



        enter image description hereenter image description here






        share|improve this answer














        I ended up grabbing the latest version of Ubuntu Gnome and reinstalling (preserving the encrypted /home).



        Additional Drivers showed me only one nvidia driver this time: nvidia (proprietary, tested) which I used. As far as I can see the GPU seems to be working as the NVIDIA settings shows more options I didn't have before and the Performance mode is now selected. I added screenshots below.



        I tested switching between Performance and Power Saving modes and it works however I need to reboot the system counter to logging out and back in as the NVIDIA settings says.



        Also, battery life is down drastically from 5-6 hours on Power Saving mode to 2.5 hours on Performance mode. This is expected.



        enter image description hereenter image description here







        share|improve this answer














        share|improve this answer



        share|improve this answer








        edited Apr 22 at 16:23

























        answered Apr 22 at 13:23









        user10853

        4711423




        4711423



























             

            draft saved


            draft discarded















































             


            draft saved


            draft discarded














            StackExchange.ready(
            function ()
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2faskubuntu.com%2fquestions%2f1026762%2fstuck-at-login-nvidia-purged%23new-answer', 'question_page');

            );

            Post as a guest













































































            Popular posts from this blog

            pylint3 and pip3 broken

            Missing snmpget and snmpwalk

            How to enroll fingerprints to Ubuntu 17.10 with VFS491