Boot fails or freezes after last update

EldrinSMP

Hello everyone. Let me start out by saying that I am relatively new to Linux and Solus, so this is my first time experiencing this type of issue. After the last update (applied for me on 2/23/24 and updating my system to Solus KDE 4.5 (6.6.17-277) I can no longer boot into Solus. Upon boot up, it shows my boot screen and then just sits there. No errors are thrown that I can see, though they may be obscured my the boot screen. When entering into the bootloader menu via F12, I have the option to boot using 4.5 (6.6.16-275) configuration instead. Choosing that gives me a blank screen but the same results. Frozen boot process with unfortunately no errors or even progress that I can report that just sits there, the current record is about 3 hours waiting, just in case. I'm really at a loss as to what to do here to rectify the situation. My initial instinct coming from Windows is to reinstall the operating system, but I know Linux is a different animal and that's a drastic option.

Some potential causes / exacerbations: Before the update and reboot that led to my current situation, I had just installed the Nvidia Proprietary drivers for my graphics card. I had also install the HP Printer/Scanner drivers. As far as I can recall, I don't think any other major changes were made to the system other than the update done through the GUI for the package manager.

System Information:
Sadly, I can't access my system to get you everything, but here's what I can remember off the top of my head.
MSI GF65 THIN series laptop
Intel i5 processor
Nvidia GX3000 series video card
32GB RAM

I apologize for not being more detailed, but unfortunately without being able to boot I'm cut off from most of that information. Let me know if you have any questions, I'll gladly answer them to the best of my ability and PLEASE help me get this thing back up and running! I am falling in love with Linux and Solus in particular, and really want to keep learning. Thanks!

Staudey

EldrinSMP I had just installed the Nvidia Proprietary drivers for my graphics card.

How did you install them? Using the "Hardware Drivers" utility?

I think the safest bet would be to remove the nvidia drivers again for now and see whether the system boots.

Try pressing CTRL+ALT+F2 (to switch to tty2), you should hopefully get a login prompt. After logging in, enter eopkg li | grep nvidia to list all nvidia packages you have install. Next use sudo eopkg rm [packagename] to uninstall all of these packages that don't have modaliases in their name (should be 2 or 3, depending on whether you've installed the 32bit driver). After that use rebootto restart your PC.

[edit: better to first try the command Reilly suggested to see whether there are issues with the kernel update]
Then when your system boots use sudo eopkg up in a terminal to do a full update (in case something went wrong there last time) and use the "Hardware Drivers" utility to install the Nvidia drivers.

alfisya

Please try these instructions here : https://help.getsol.us/docs/user/troubleshooting/boot-rescue/. Let us know if it helps.

EldrinSMP

alfisya Thank you for the link. As I do not multiboot (when I switched to linux from windows I decided to jump into the deep end and abandoned a Windows installation completely, thus my extremely rudimentary understanding of the OS and terminology) and have not seen any emergency prompts, I'm working on the assumption that there is an update issue. Following the instructions in that section, I run into a problem when trying to get networking working in the "chroot /terminal" environment.

When running the "ln -s ../run/systemd/..." command I get the response:

ln: failed to create symbolic link '/target/etc/resolv.conf': File exists

Removing that file and recreating it with the before mentioned command is successful, however when I return to chroot I still do not have network access. As you can guess, this causes failure with the eopkg up and subsequent commands.

Searching for broken packages DID return an interesting result. The package for my python3 installation (pre-installed Python 3.10 installation) seems to be broken. I can't fix it due to networking issues, but could that be the cause of my failure to boot?

ReillyBrogan

Can you switch to a tty from there? Also, if you tap the down key you should be able to show the boot menu. Hit e on the kernel to edit the command line arguments and then delete quiet splash systemd.show_status=false and hit enter.

EldrinSMP

ReillyBrogan Thank you for your reply. I apologize but I'm just starting to learn Linux and Solus, and I honestly don't know what you are trying to ask me. Could you please rephrase like you're talking to a complete newb? LOL

ReillyBrogan

EldrinSMP When you turn on the computer rapidly hit the down key. A kernel selection menu will appear (if it doesn't show up you may need to try again, the window to get into it is fairly short). Find the kernel 6.6.17-277 and hit the e key while it's selected. This will allow you to edit some of the command line arguments used for starting the kernel. Find the arguments I mentioned and delete them. Then hit enter to boot. This will cause a lot of things that are typically hidden during boot to be shown, you can take some pictures of this with your phone whenever you see something that looks like an error and post them here.

EldrinSMP

ReillyBrogan ReillyBrogan Perfect! Thank you!'

Actually, almost everything runs and completes like it should, then it just... sits there. This is what is currently displayed on the screen. It's been this way for about 10 minutes now.

ReillyBrogan

If you hit CTRL + ALT + F3 does it switch to a terminal? If so login and run sudo clr-boot-manager update, see if you get an error.

EldrinSMP

ReillyBrogan Everything I learn about Linux I love it more and more. Yes, I was able to get to a terminal and log in using the TTY command. I ran the command you suggested, and got no errors.

Staudey Yes, I installed the drivers via the "Hardware Drivers" utility. Until I get more experience, I'm relying on the built in systems as much as I can. I was able to get to a terminal, and removed the NVidia drivers. You were correct in assuming they were the culprit. I am happy to announce I am back in my system and everything is working as intended. I think I'm going to wait until I actually have a need to try reinstalling the Nvidia drivers. I was HOPING they would help with the performance of the GNOME Network Share Miracast app I was testing to be able to get my wireless second display back and actually usable, but if they're going to cause this kind of trouble, I can live without. Thank you very much for your help.

To everyone that stepped forward to help me, I want to say thanks. You guys never once made me feel like an idiot for being new to this, and I've never encountered that before when seeking help online. You're all awesome!

ReillyBrogan

EldrinSMP So the Nvidia drivers should have worked, could you tell me what you get when you run uname -r?

EldrinSMP

ReillyBrogan I'm really sorry, I missed the notification that you responded. When I run the uname command it returns 6.6.17-227.current. If I'm not mistaken, that's the Linux Kernel version? Isn't that something that gets updated through a Solus version update?

Also, I should have included this earlier, but my graphics card is an Nvidia GeForce RTX 3060 Laptop GPU... Could the drivers we have just be incompatible with that model/series?

Staudey

EldrinSMP If I'm not mistaken, that's the Linux Kernel version? Isn't that something that gets updated through a Solus version update?

Yes, it gets upgraded with a regular system update. It's just that in edge cases, like when your system has too small an EFI partition, while the kernel package gets updated, it might not actually apply to your bootloader (that's why really suggested running sudo clr-boot-manager update and seeing if there are any errors.)

I think you have a typo there, 6.6.17-227 shouldn't exist. Is it maybe actually 6.6.17-277? If so, that is fine and you are on the newest kernel.

EldrinSMP Also, I should have included this earlier, but my graphics card is an Nvidia GeForce RTX 3060 Laptop GPU... Could the drivers we have just be incompatible with that model/series?

The drivers should be compatible with that GPU. A potential mismatch with the kernel version would lead to problems with the nvidia driver though. But I think we can rule that out now.
Could it be that your system was not fully updated last time?

EldrinSMP

Staudey

You are correct, it was 277 not 227. MY bad. Regarding updates: I ran the sudo clr-boot-manager update command and it goers staight back to the terminal prompt so I assume there are no errors. I've run 'sudo eopkg up' as well. Everything is up to date. I've installed the drivers again using Doflicky and it get the same result. It seems to crash just as the X Server is activating. I've tried uninstalling the xorg-driver-video-nouevau driver, same result.

I currently have three working theories and various levels of ability to test them. The first is that I'm beginning to think that it's a problem with the driver itself. I've been researching the issue on other distros, and the 535.* and 545.* series of drivers caused a LOT of problems, especially with the RTX 3060 that I have and one of the most common is exactly what I'm experiencing. However, the beta driver (official driver as of two days ago) is 550.54.14. The BETA on the repository still has the issue, I wonder if the production one would once we get it in. I'm considering trying the older 470.* driver as well.

Second theory is that it might have something to do with my onboard video card, the Intel Corporation CometLake-H GT2. The 'lspci -k | grep -EA3 'VGA|3D|Display' command gives me the following:

With the Intel as the primary, is it possible that something done during the installation process of the NVIDIA drivers are causing the Intel driver not to load, causing the freeze. This was a proposed root cause of the issue in a few Ubuntu forums. I can find no xorg config or blacklist files preventing it, however, I am also very limited to the command line when the drivers are installed... How can I check this?

My final theory is that it might be a signing issue with the UFEI Secure Boot. If the Nvidia drivers aren't being signed and registered like they should, that could be causing the issue (if Solus even works like that... It was the root cause of a number of Unbuntu issues with the same scenario). If that's the case, it should be an easy enough fix, I just have to generate a key, etc. etc. etc.

I will admit, I'm getting frustrated. If I could figure out how to kill the X Server and a myriad of other things, I would install the package offered directly by Nvidia, however the setup of those makes this troubleshooting feel like a cool spring breeze on a hot day. LOL