So my system is not booting correctly and I'm pretty much at wit's end. The problem arose when I tried to get CUDA 9.2 (which I need for a dependency related to PyTorch) to work with my Ubuntu 18.04 LTS distribution. I know these are not technically supported together, but using the first answer to this question, I actually got it to work and the system did boot successfully at least once after that.
The first time I had this problem, I was able to solve it pretty easily by purging everything related to the NVIDIA drivers and falling back to the default graphics driver installed on the system (nouveau). The problem quickly reappeared, however, this time the GUI crashed while I was trying to install CUDA. My previous solution didn't work.
I can get to the GRUB bootloader and start the system in recovery mode, where I can access a shell. Some of the tty-x screens are also usually working. I ran
sudo apt remove --purge *nvidia*
sudo apt autoremove
dpkg -l | grep nvidia does not return anything. The same command with nouveau yields
ii libdrm-nouveau2:amd64 2.4.101-2~18.04.1 ...
ii libdrm-nouveau2:i386 2.4.101-2~18.04.1 ...
ii nouveau-firmware 20091212-0ubuntu1 ...
ii xserver-xorg-video-nouveau 1:1.0.15-2 ...
Currently startup gets stuck at
/dev/sda5: clean, 1806248/12214272 files, 36749798/48827904 blocks
_
I would appreciate any ideas, if the output to any specific commands is useful to debug this please tell me and I will provide them. Please note that I'm relatively new to Ubuntu (and also to dealing with CUDA), so I don't really know which additional information would be useful.
Thank you in advance,
Greetings.
mountcommand did not seem to do anything, and the other commands I have already tried (NVIDIA was already purged,ubuntu-desktopand its dependencies already installed). – Jonah Mevert Oct 08 '20 at 13:51