[SOLVED]Arch freezing on loading initial ramdisk with four purple dots

Amasirat · 2023-09-02 10:54:05

System:

Motherboard: Gigabyte H110M-S2PT
CPU: Intel Core i5-6400, (4 core)
RAM: 8GB
GPU: Nvidia Geforce GT 740
Hard Drive: HDD 500GB
DE: KDE Plasma

As detailed as possible story leading up to the issue:

I apologize if the extra details are unnecessary however I'm putting as much detail as I can so that the issue becomes clear.

My computer was a dual boot windows+Archlinux setup. For whatever reason I couldn't setup chainloading with Grub, so I just let them use different boot loaders. That worked and I only had to change boot device inside the motherboard's BIOS/UEFI menu.
One day, I installed nvidia-470xx-dkms from the AUR which turned out fine I had believed. I remember dkms installing it correctly. in that same session, the system froze completely which gave me no other option than hard reboot.
the freezing had happened sometimes however I couldn't deduce the reason at those times.

This time, the computer freezes when the kernel is loading. specifically when "loading initial ramdisk" is printed on screen. with four purple dots right below it. after lingering on that screen for a while, a few stripes of corrupted multi-colored pixels show up on screen. (I'll upload images of my monitor I took from my camera phone below)

I thought I could access the system through the live environment, but the same thing happens when booting that.
At last, I tried using another distro, namely lubuntu, however the same thing happens with that as well.
The only OS that boots up is Windows 10, however windows 10's resolution is widened.

Fortunately, I don't have much critical files stored on my computer that I require, because I'm very careful with storing backups of my files and most of my important projects are on github, So I could find a way to wipe the entire drive and start over however this issue is so peculiar and strange to me that my instincts tell me I have to understand what it is first before doing anything else.

I put this on here because it appears to be a kernel issue since there was first a kernel panic and then loading the kernel is the thing that's not successfully accomplished. It could be anything else however I'm no expert. Feel free to inform as to how to categorize this issue.

I have put every detail I believe that might have been the cause there.
my guess suggests it could be:

a) The nvidia-470xx-dkms and its various sub packages that I installed like its lib32 version or opencl, usually a failure of that would cause only the xorg server to not be able to run so I see this as unlikely.

b) The neglect to setup proper chain loading with Windows10 on Grub.

c) It could be that my HDD has been corrupted in some way, It is rather old after all, Plus I had run into some forums which discussed a hard drive fault as a cause of regular system freeze, however I could not verify if this was indeed the case for me.

I guess right now all I require is to be able to access a log or something, so that I can pursue this issue further. Any help with understanding this issue is appreciated.

pictures of how system behaves while booting kernel:
https://imgur.com/a/N5ewFdt

EDIT: Big surprise, it was the Nvidia driver. Shocker I know. The reason was that the noveau module had conflicts with nvidia-470xx-utils. I made a mistake and didn't remove kms from the mkinitcpio config file. Moral of the story: READ ARCH WIKI BETTER

Last edited by Amasirat (2023-09-02 20:30:38)

seth · 2023-09-02 15:11:25

https://wiki.archlinux.org/title/GRUB/T … ramebuffer along maybe "nomodeset" will hopefully allow you to boot the multi-user.target (2nd link below) and from there save data as well as extract some information out of the system (notably about the condition of the GPU)

Don't reboot w/ the powe button, use https://wiki.archlinux.org/title/Keyboa … el_(SysRq)
This way you'll hopefully preserve a log of a bad boot.

Do you have a spare GPU?
Resp. the CPU should have a HD Graphics 530 IGP, so removing the GPU might allow you to boot w/ the intel chip?

Amasirat · 2023-09-02 19:38:58

Hey! It's good seeing you again sir!
You were very patient with me last time when I had that network issue, so I wanted to say thank you for that.

Unfortunately no spare GPUs are available to me at this time. However I managed to boot into system by setting nomodeset quiet splash parameter. after checking the journalctl logs of the previous boots it seems it actually was an Nvidia issue yet again. I swear that nvidia-470xx package has caused no shortage of headaches for me, well my legacy nvidia GPU is the cause of it in the first place.

NVRM: The NVIDIA probe routine was not called for 1 device(s)
NVRM: This can occur when a driver such as:
NVRM: nouvea, rivafb, nvidiafb, rivatv
NVRM: was loaded and obtained ownership of NVIDIA device(s)
NVRM: reconfigure your kernel without the conflicting
NVRM: driver(s), then try loading the NVIDIA kernel module again
NVRM: No NVIDIA devices probed.

It seems from these logs that the proprietery driver did not blacklist the nouveau module correctly, however I should be able to take it from here.

I was not aware of nomodeset and sysrq especially. Thank you again for your assistance.

EDIT: By the way, is editing the title to contain [SOLVED] enough?

Last edited by Amasirat (2023-09-02 19:40:09)

seth · 2023-09-02 20:07:47

https://wiki.archlinux.org/title/NVIDIA#Installation - pay attention to #5
Once this is fixed, yes please add "[SOLVED]" to the title.

Arch Linux

#1 2023-09-02 10:54:05