Recently, after updating my system, it no longer boots. I tried rolling back packages through chroot on a live USB and the rollback failed because youtube-dl no longer exists (replaced by yt-dlc). How can I roll back my system? Alternatively, if there is another way to make my system bootable, please let me know.
stdout: https://0bin.net/paste/G0Dnr1ZG#kzrR-j/boebr0eHp8OW9AdRUJoT43NQwEsChWGt4pEn
stderr: https://0bin.net/paste/hEB9f-rS#erNU8giRpo41ZSWa8jy6aCRel4RJFnUke7ZA28Jkmrq

    alex5nader this has happened to me in the rollback process before (looking a deprecated package) but I can't remember if I got around it.

    What's weird in the first pic is the Y/N for youtube-dl. I don't remember one for me, maybe it just plowed ahead without me. spend some time in live USB eliminating your variables.
    If you get the tty you are halfway there. sounds like something got mangled in the install process.
    Run the whole checklist solus gives in Main Site/Troubleshooting/Booting (or something).
    This has happened to me a couple times and it sucks. I'm tempted to say your DM got divorced but there could be a few reasons this happened..
    edit: this is from today. maybe applicable to you? https://discuss.getsol.us/d/5689-black-desktop-after-today-s-update/12

      Seems I misunderstood what was happening. I am able to get to a tty via ctrl+alt, so it is booting. Switching back to ctrl+alt+7 just leaves a black screen with a blinking cursor. It doesn’t seem to respond to any input. I’m running Budgie, not Plasma, so my issue is probably different from the one you linked. I went through the boot recovery page, and it didn’t seem to help. Let me know if there’s anything else I should try.

      after tty:
      sudo eopkg check - to see right way what may be corrupted
      sudo eopkg check | grep Broken | awk '{print $4}' | xargs sudo eopkg it --reinstall --if you want to check and fix at the same time

      then try update again
      sudo eopkg up

      there's a few more things but that's the basic package check

      Those were on the boot recovery page, so I don’t think they helped. I did run them from inside chroot, rather than on the actual system, though I don’t see a reason for them to work differently between chroot and the actual system.

      I'm making food and won't be back a while. Use search function here for LIGHTDM. maybe you have to restart the service from the tty. or similar.
      good luck, hopefully another forum member pitches in--

      so what you are saying is all your packages are good then? you saw and OK next to everything? just chekin

      I found this in journalctl, which looks suspicious to me:
      NVRM: API mismatch: the client had the version 455.38, but
      NVRM: this kernel module has the version 455.28. Please
      NVRM: make sure that this kernel module and all NVIDIA driver
      NVRM: components have the same version.
      My nvidia drivers were updated between those versions during the update which broke my system. How can I upgrade the kernel module or downgrade my drivers? Should I even do either of those?

        alex5nader Also, lightdm is exiting with status=1/FAILURE. Additionally, Xorg.0.log notes that the nvidia kernel module failed to initialize, and then X also exits with error (1).

          yay! sauce is cooling.
          not yay! this happened to me in march. kaput. something updated (I imagine) triggered by noveou drivers to tank thus divorcing lightdm and no real OS.
          you did great detective work.
          both those warnings related.
          somewhere (https://discuss.getsol.us/d/1675-nvidia-proprietary-drivers-confusion/3) you have to figure out if you ave kernel compatibility.
          I was not that patient. My reading---and I'm sorry I NEVER write my fixes down because I just want to get up and running---led me to believe I needed the nvidia 390's for my Geforce board. i installed them in the tty then had to manually fix the lightdm service restart then bam I was whole again...
          follow more threads I think you are really close
          sorry I wish I had the magic instructions. this is not real common but not unheard of; it happened to me

          Installing 390 did change the client version, but the kernel module version was still stuck at 455.28. It seems that for some reason I’m unable to change the kernel module version.

            alex5nader I guessed and got lucky. I wouldn't do that unless you know your mobo/graphics matched the kernel. I got "lucky" (if that' what you call 1.5 days to restore sys.) in that the move I made was compatible with the kernel and I was too dumb to know it. As I said, had to guess based on reading it was the 390. If my sys was already broke I did not think I could break it more.
            and work thru the live usb
            and do baby steps in the tty
            ---quite a few threads about the closed source nouveas (via solus' doflicky) saving the show instead of NVIDIA so I'm at the edge of my knowledge

            there's some nvidia compatibily experts around here, I hope they chime in. still think you are close but I'm out of advice except read read read and you are doing that.

              I'm not dual booting, and I had the proprietary drivers (nvidia-glx-driver-current) installed and working before the upgrade. Those posts were nice in understanding the differences between the different nvidia packages, but I'm still not sure what the difference is between the "client" and the "kernel module", or how I can install one vs the other. I thought the nvidia-* packages are intended to include both, so I'm not sure why the kernel module is not being upgraded like the client is. For good measure, I'm using a GTX 1070.

                alex5nader
                What caused your initial problem I have no idea. But the rollback failing midway due to the youtube-dl rename would explain the missmatch (actually it fails before installing ignore that). So would choosing to boot into the older kernel with an up to date system.

                Check the output of uname -r
                We should be on 5.6.19-159

                eopkg info nvidia-glx-driver-current should be

                nvidia-glx-driver-current, version: 455.38, release: 354

                So you're better off moving forward again sudo eopkg up -y and troubleshooting the initial issue because you can not go back, not fully anyway.

                    I don't think the rollback actually applied any changes since it failed partway through. Upgrading again finds no changes to make. Also, this nvidia driver issue was in journalctl from after the upgrade but before the rollback, so I believe it is the original problem, and not something left over from a bad rollback.

                    Edit: are you able to rename this post's title to "Nvidia kernel module not updating with package"? I can't seem to edit it myself.

                    Harvey changed the title to Nvidia kernel module not updating with package .

                    Harvey nvidia version is correct, uname -r gives 5.6.19-158.current.

                      Yeah then either you're not up to date or its failing to install the new kernel image correctly. This usually happens when the UEFI partition is too small. But your not dual booting so that becomes less likely unless you did custom partitioning when you installed Solus.

                      sudo eopkg it --reinstall linux-current
                      See what errors if any it gives. If none try rebooting again.