Weird kernel parameters?
From the ISO I assumeâŚ
1 Like
Did you use the proprietary Nvidia drivers when you used EndeavourOS for the first time?
And do you remember how you installed at that time?
So far the nouveau open source driver looks like the only working solution.
And sorry that I canât really help more, but maybe you could look at the Arch wiki and the EndeavourOS wiki to see if they have any solutions for your machine.
@rs265
Have you tried installing the nvidia drivers manually?
sudo pacman -S nvidia nvidia-utils lib32-nvidia-libgl nvidia-settings
I do find the issue with btrfs a little puzzling since i use it and Iâve tested so many times without issue. Are you using erase disc and swap file with btrfs as the file system or are you doing some manual partitioning for the install because it doesnât make sense to me that itâs failing.
I know that I had nvidia on my previous iteration of EndeavourOS which was my stable distro for a good year or so. I have been distrohopping for about 3 months. I donât remember the version of the Nvidia driver I got up to that time.
Since we havenât gotten anywhere with these attempts (no offense, thanks for all the suggestions so far), I am going to attempt a reinstall. I am wondering if I should use an older version of the installer and see if that installs with nvidia and then an update to the latest versions works from there.
I guess you could try an older iso and see if it boots on the nvidia option in the menu.
The old ISO might not work anymore.
But if you reinstall, try also using directly the proprietary option if the open source option fails (or vice versa 
).
Or try installing the offline version of KDE.
Manual partitioning scheme. I have pre-defined partitions from root, home, boot, efi and swap. So, when I reinstall a fresh system, I usually format my root, boot, efi and swap and keep the home stable on ext4. efi has to be fat32 and swap is just swap. I tend to avoid messing with boot partition too much and leave that at ext4. The only one I play with is the root partition and every attempt with btrfs leads to a rescue mode and utter failure. I have always wanted to try btrfs but it apparently has it in for me!
Well, I will begin my next re-install attempt nowâŚ
I never should have given up my year of stable EndeavourOS to go distrohopping all primarily to use an âofficially supportedâ VPN application just to find the application gave me nothing more than downloading a bunch of conf files from the account or the fact that the app is actually available in AUR and wasnât showing for me when I searched for it!
As for the offline KDE, I will have to go that route if I try an older ISO especially if I can get the NVIDIA option at boot to work on it. Then, hopefully an update would work from there!
Good luck! If the old ISO works, note that it is possible to install another DE after the offline install (but be sure to run eos-update right first reboot).
Well there must be something up with the manual partitioning method as the btrfs setup automatically creates the subvolumes on the install. Iâm not sure how you could save /home since it already sets up a subvolume for that. I am using grub btw not systemd-boot. Here is how eos creates it using erase disc with swap file.
âânvme0n1p1 259:2 0 1G 0 part /boot/efi
âânvme0n1p2 259:3 0 464.8G 0 part /var/log
/var/cache
/swap
/home
/
This explains why btrfs fails with my manual partitioning setup then. Is there no way to tell btrfs to limit itself only to the root partition? Also, is grub necessary? I was a grub person for many years before switching to systemd-boot - find it to be simpler and quicker to load at startup.
I created an ISO from Galileo build of Nov 2023. This uses NVIDIA v545.xx drivers. The NVIDIA boot option on the live USB works and the installed offline KDE system with 6.6 kernel worked fine.
Then, eos-update happens and NVIDIA moves to v565.xx. There goes my graphical screen! I tried to get the update to skip nvidia packages but couldnât get that to work for some reason. So, now I need help to see if downgrading/upgrading the NVIDIA drivers fixes my problem. My working theory is that this particular version of the driver is not playing nice with my system. Please help!
You could try the LTS kernel and use that for now:
sudo pacman -Syu linux-lts linux-lts-headers
linux-lts is version 6.6.64-1 currently.
If you have package nvidia-dkms, then no other package is needed.
But if you have just nvidia, youâll need to install nvidia-lts too.
I use grub because i use snapper to take snapshots and with grub you can boot on a snapshot and then restore. I donât think you can do that with systemd-boot. I also find grub works as quick and is just as simple.
You could have used eos-shifttime to go back.
1 Like
I am on a fully updated system and have added the linux-lts as per your suggestion. Both 6.12 and 6.6 kernel have the same blank screen behavior with NVIDIA 565 drivers. I need help getting the graphics back to functional. Thanks.
tried eos-shifttime but that requires graphics or something else that my system is incapable of providing at the moment!
I need a way to get nvidia to work. I tried a downgrade method suggested in the last post on How can I roll back to the previous version Nvidia? - #8 by manuel but that only got me as far as the most recent older point release. Going further back gave me some library conflict between nvidia-utils and egl-gbm or something like that. Please help! After 3 days of this, I am reconsidering distrohopping again!
Sorry, Iâm out of ideas. Hope someone here can provide more help.