EndeavourOS boot stuck after forcefully shutting it down

I was playing around with my KVMs and i faced the vendor-reset bug. With nothing else to do (that i know of), i force restarted it by the button on my computer case. I have done it about 10 times before but this time, i think it broke my system.

I have snapper and btrfs installed and configured so that when things like this happen, i can safely return to my systems. But booting to a snapshot gives me the same error. Yes i have tried different snapshots. The error messages shown were:

[FAILED] Failed to start **Record System Boot/Shutdown in UTMP**
[FAILED] Failed to start **Take snapper snapshot of root on boot**
[FAILED] Failed to start **Regenerate grub-btrfs.cfg**

Now, i just get stuck here. Seeing this, i think it might be snapper that is being a bitch. Any ideas what i could do to fix it?

snapshots live in the same volume as the original data so disk corruption can impact them as well.

I would boot off a live ISO and check the integrity of your btrfs volume.

You should also setup the magic restore key bindings as well.

i booted off of the endeavourOS live cd and ran this:

btrfs check /dev/sda4
btrfs check --clear-space-cache v2 /dev/sda4

sda4 being my root partition. after doing those, i get the same error when booting up my main system. so that didn’t work.

You should also setup the magic restore key bindings as well.

@fbodymechanic yes i already did that. what would i do with it?

Well, you should try to use that instead of a hard power off.

Can you explain what you mean with these two?

I was playing around with my KVMs

KVM is a kernel virtual machine, a type 2 hypervisor. It is a virtual machine, completely isolated from my main endeavourOS system. I normally use Windows on a virtual machine because of Microsoft’s aggressive telemetry and/or any type of shady programs i need to run. Plus another KVM for gaming.

i faced the vendor-reset bug

The vendor reset bug is a known bug on AMD graphics card. So if you have a KVM setup with only one graphics card on your system, linux has to take the graphics card off of itself and pass it on to the windows virtual machine. we have to make scripts for that detach/attach. My scripts work fine, as they both allow me to get inside the Windows virtual machine with full graphical acceleration and get back to my linux host when i shut my Windows virtual machine down. That should’ve been the end of it. However, I use an AMD graphics card. AMD cards have this bug that gives you a black screen when you reboot your Windows virtual machine. As i didn’t reboot that much, i postponed fixing it for a while. I finally got around to applying the patch from AUR before my system faced that problem with snapper. (i think that’s not the source of the issue as i have rebooted several times after installing that.) I turned on my Windows virtual machine to test if i applied the vendor-reset patch correctly and i did not so i had to reboot forcefully. That’s when the problem happened.

I suppose you know, but I have to note that virtual machine files would require special configuration with btrfs system. Do you confirm you have configured KVM properly in that regard?

Other than that, my only idea is to chroot to your system and troubleshoot, recreate kernel images and other usual stuff. :man_shrugging: