I confirm, everything works very well. I installed kmod-open-desktop, works smoothly without errors. Thanks for your interest and quick solution to the problem. You guys are awesome!
Ah, no it was just coincidence - SDDM & nVidia problem still exists when I am using kernel-desktop, if I run kernel-server I can use SDDM and login directly and nvidia driver is not crashing anymore
I am on X11
Will be trying out the 570 drivers once I am able to. On my end it is saying that I am missing libcrypto.se.1.1()(63bit) and egl-wayland. It has also happened on Linux Mint once in a while, but the solution in those cases was to find those in the console with custom repositories and install them through there. I guess these issues will be fixed once I am able to update the mirrors again?
I was wondering though why the 565 drivers will no longer be available now when the 570 are there. Isnāt it better to provide options for those who want to try different numbered drivers depending on if the latests drivers happen to bug out their system? Pretty unfamiliar with rolling distros and if this is considered standard to only offer the latest drivers. Just want to be more informed.
I tried the dnf distro-sync again and it updated a number of packages. I was then able to install ānvidiaā cleanly, getting the 570 version of the drivers. But now I canāt boot. No matter what I choose at the bootloader, it hangs on āLoading initial ramdisk ā¦ā.
I have three sets of choices in the boot menu:
Linux 6-14.0-desktop-0.rc3.1omv2590
Linux 6.13.4-desktop-2omv2590
Linux 6.13-9-desktop-1omv2490
None of the āconsole modesā get anywhere. I tried the ārecovery modeā and that hangs at the same place too. I know I can use dracut to regenerate the ramdisk, but I canāt get to a console to run the command. It seems you canāt ctrl+alt+f4 to a text console from there either.
How do I fix this, and how did I end up here?
Thanks.
@OhPlz are you booting with 2 monitors plugged in?
No, just one monitor using display port.
Given that, itās hard to even investigate the issue.
The only thing I can think of is one of these methods might help, however you need someone experienced to guide, if the case:
Resources Index
Top spot - How to fix broken boot-loader
Top spot - How to use āLiveā OMLx iso to repair broken OMLx system
I am wondering if the install of the Nvidia driver did not run the code to rebuild the initramfs. If you can boot grub to a command line, try running this.
sudo dracut -f /boot/initrd-6.12.9*desktop*.img $(uname -r)
If you login as yourself, use the sudo. If you login as root, you wonāt need it, of course.
Thatās what Iām thinking too, something neglected to rebuild the initramfs, although itās curious that booting an older version doesnāt help. I donāt know if thatās expected or not, Iām a little out of my depth on boot issues. The only command line I can get to is the grub command line, which isnāt a linux shell.
Looking at the grub entries, the newest ā6.14.0-desktop-0.rc3.1omv2590ā has initrd set to ā/boot/initrd-6.14.0-desktop-0.rc3.1omv2590.imgā. I think the 6.12.9 would want to be 6.14.0 in the dracut command that you posted⦠if I could get to a prompt that could run dracut.
I might be looking at a reinstall here, and then try to get back to that point and run dracut right after installing nvidia. Unless someone has an idea on how I can to a command prompt without needing to reinstall. Could I āchrootā from a command line from the Live USB and then do that?
You could probably chroot to it, but I doubt I can walk you through that.
I was hoping grub would give you one option to boot to a command line.
I tried fixing it, but only made things worse. Ended up reinstalling with a freshly downloaded Live USB. This went much more smoothly. Iām now happily running 570.86.16 of the nvidia driver. The only problem I did run into was that KDEās power management service started crashing afterwards, shortly after logging in. I was able to fix that by adding:
POWERDEVIL_NO_DDCUTIL=1
To the ā/etc/environmentā file.
It has something to do with setting the brightness of the monitor and is apparently known to be flaky. Havenāt seen the issue since.
Did you install nvidia-settings?
Only ānvidiaā and ānvtopā.
I did just install ānvidia-settingsā. Thanks for mentioning it, I had totally forgotten about it.
I did get bounced back to a login screen right after logging in today after a cold start, followed quickly by a black screen with a blinking cursor, so maybe thereās still some problem lurking. Nothing obvious in journalctl, and there was no powerdevil crash. Not sure what thatās about. Was able to login after rebooting.
EDIT:
Wild⦠sounds similar to this brand new thread: Black Screen with blinking cursor apears 30 seconds after the desktop loads
Try to install and run kernel-server instead of kernel-desktop - for me it is fixing this nvidia problem in 90% of boots.
I have a stranger work-around than that. If Iām starting from a cold boot and simply wait a couple of minutes at the login screen, I can then enter my password and login and thereās no issues. Makes me think thereās still something initializing in the background when the login screen is first presented.
Thatās possible.
Actually, after last kernel / nvidia driver update in ROME this solution is less viable now.
In my tests 40% of logins with kernel-server are successful, 60% crashing after 1 min.
On kernel-desktop around 10% success, 90% crashing.
So I have to back to:
CTRL+ALT+F4, login in console
sudo systemctl stop sddm
startx
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.