Hi, I’m pretty new to Fedora (and Linux in general) and upgrading from Fedora 41 to 42 is my first in place upgrade, so apologies in advance if my questions here are trivial. I’ve been thoroughly enjoying using and learning about Fedora but can’t figure out what I messed up in the upgrade.
TLDR: Please help, all of my boot options give me a blank screen with a useless cursor after trying to upgrade to Fedora 42.
I initially installed the upgrade from the “Updates” section of the app store after installing app updates, and I got an update screen on reboot with a progress bar that eventually completed. However, when starting again I was dropped into the BIOS menu, showing 4 options:
Fedora Linux (6.14.3-300.fc42.x86_64) 42 (Workstation Edition)
Fedora Linux (6.14.2-300.fc42.x86_64) 42 (Workstation Edition)
Fedora Linux (6.13.11-200.fc41.x86_64) 41 (Workstation Edition)
Fedora Linux (0-rescue-<long string>) 41 (Workstation Edition)
I was able to boot using the 3rd option (41) successfully, and attempted to upgrade by following the docs. After running
sudo dnf upgrade --refresh
I tried to reboot like it says, but when I selected the 41 option again, I got a wall of logs (all starting with OK) and then a black screen with a blinking cursor at the top left. Selecting either of the Fedora 42 options from BIOS gives me the same result. I can type at the cursor but it doesn’t seem to do anything and gets cleared after 10 seconds or so. Is this common when upgrading? I’ve seen some stuff about Nvidia drivers causing issues but not sure if this is the problem? Any help or advice would be greatly appreciated and thank you in advance!
Computer specs:
Lenovo Legion T5
Intel i5-10400
Nvidia 1660 Super (I installed drivers following these docs)
Running Workstation and Wayland if that makes a difference
Disk is not encrypted
Other things I’ve tried:
Selecting the rescue option from the BIOS menu repeatedly gives me the message “Cannot open access to console, the root account is locked”
Adding/removing the rhgb quiet,rd.driver.blacklist=nouveau, and nomodeset=0 kernel parameters hasn’t changed the outcome
Hey thanks for your reply! I’m away from the machine at the moment but I’ll try both of these things today. I don’t think I’ve tried just nomodeset, will see if that makes a difference. I’ll take a closer look into single-user mode, but it seems like that should let me use journalctl at least.
Update: nomodeset without =0 doesn’t seem to make a difference, and adding single to the kernel parameters doesn’t seem to change anything unfortunately. I tried this with the rescue option and other ones and still get the same results so far.
please post grep CMDLINE /etc/default/grub and dnf list \*nvidia\* --installed
Can you boot successfuly the newest kernel with nouveau driver instead of the nvidia driver? Boot without kernel parameters rd.driver.blacklist=nouveau modprobe.blacklist=nouveau
EDIT:
rescue kernel: when did you refresh the rescue kernel the last time? file /boot/vmlinuz-0-rescue* shows the kernel version.
This worked, thanks so much!! I was able to boot using the nouveau driver by removing the rd.driver.blacklist=nouveau kernel parameter.
To answer your questions: grep CMDLINE /etc/default/grub gives me GRUB_CMDLINE_LINUX="rhgb quiet rd.driver.blacklist=nouveau", which matches what I saw when editing the kernel parameters from the BIOS menu so I think that makes sense. dnf list \*nvidia\* --installed gives me a list of the akmod, kmod, and xorg Nvidia drivers that I installed. I’m thinking that I may have to remove these and install updated packages because the software app is telling me I have updates to download but giving me file conflict errors with my Nvidia packages as well. file /boot/vmlinuz-0-rescue* is telling me that it’s been a while… I guess it’s time to figure out how to update this for next time.
Out of curiosity, if I hadn’t been able to boot with the modified kernel parameters, what would have been the best way to run these commands? Would I need to boot to an external drive? I didn’t find the GRUB command line helpful, but I’m definitely a noob here and may have totally missed something there as well. In any case, thanks again for your help, I really appreciate it!
this was not a solution. It shows us that something is not right with the nvidia driver on your system.
please post the actual output of dnf list \*nvidia\* --installed as preformatted text ‘</>’
it looks like the kernel parameter modprobe.blacklist=nouveau is missing on your system.
try this:
boot the newest kernel with nouveau driver activated
Sorry about that, I figured I’d try reinstalling the drivers but haven’t had a chance to take a look yet. After messing with this for a bit, I’ve found that I still have the same 4 boot options, but the one labeled 41 is actually the only one that will boot (when I remove the rd.driver.blacklist=nouveau kernel parameter). However, when I boot from the 41 option, cat /etc/fedora-release gives me Fedora release 42 (Adams).
This is my output from dnf list \*nvidia\* --installed
I tried running these grubby commands on the 41 boot option (which says that the release is 42) since that’s the only one that works. This successfully modified the kernel parameters I see on boot, but the system still doesn’t successfully boot by default, it only works when I manually remove the nouveau kernel parameters.
Is says 41 because kernel was installed while the system was still running F41.
How did you upgrade from f41 to f42, by dnf system-upgrade (CLI) or a GUI tool like Gnome/software?
There is a mismatch, the kernel modules are 570.144 but the binaries are still 570.133.07.
Do you remember from which repository the nvidia drivers were installed initially? Then the package nvidia-kmod-common.arch is a hint that this is driver packaged by Negativo17
We need output of the following commands:
alias dnf
sudo alias dnf
ls -l /usr/bin/dnf*
ls -l /etc/rpm/*
dnf5 repolist
dnf5 repolist --disabled
dnf5 list \*nvidia\* --installed
cat /etc/default/grub
sudo dnf5 check-update
sudo grubby --info=ALL | grep -A 4 index
sudo dnf system-upgrade log
It seems you might have packages installed from different repositories. It will be best to remove the driver and reinstall from a single repository.
Sorry for the late response here! I originally upgraded via the Gnome software app, and then tried doing it again via CLI because I assumed something wasn’t right.
From running a few of the commands you listed, I believe I’m running dnf5 and it’s up to date. However, I was messing around trying to upgrade the drivers and ran
sudo akmods --force
sudo dracut --force
and then rebooted. I was hoping to ensure that the kernel modules and boot image were both updated, but unfortunately now none of the boot options work, regardless of the nouveau parameters that I use and I’m back to the blinking cursor. I’ll probably try a few more things and then just do a fresh install since I don’t have much that wasn’t backed up. Thank you again for the help!