Hello everyone, I’ve only been using the Linux Fedora operating system for a few months, and only use it for entertainment purposes. That’s all to say I’m still very much an amateur when it comes to understanding how Linux operates and how to properly troubleshoot it. So I would greatly appreciate if you could explain how I can fix the problem like I’m five. Thank you in advance.
The issue I’m currently experiencing is that after the latest Fedora 44 package update. I updated and restarted my computer as normal. The pc asked which kernel version I wanted to use. I picked the latest updated kernel. The Fedora loading screen shows up, but after a few seconds, it goes to a black screen with only a white blinking underscore to accompany it.
Here’s my PC and OS setup: Fedora 44 KDE Plasma Desktop Edition, Graphics Card Nvidia 660 GTX. I don’t know if this is helpful, but I’ll try to provide more information about my system if needed.
I have tried to resolve the issue by first rebooting my pc just to see if loaded incorrectly. Second try was to go back to a previous kernel to see if it was a problem with the current kernel update. All other kernels that were available to me on the boot screen did the exact same thing as the newest updated kernel. Here are the Fedora kernels that were available to me 7.0.12.201, 7.0.12.200, 7.0.11.200. All are KDE Plasma Desktop Edition. Thrid try was to boot into my command line and try to log into my desktop that way. It took a couple of times to login into the terminal, so I got a couple messages saying I failed to log in. Until finally a green line of text appeared with a dollar sign. Which I read on the forums, that it means I’m logged in. After that, I typed in sudo systemctl restart kde to restart the display manager, with a message shortly after saying it failed to restart because it could not find the service. I wasn’t sure what to do next, so on the fourth try my stepdad suggested we check the graphics card to rule it out. We swapped the card out with a similar but newer card. We then loaded in Mint, POP, and Fedora 43 to make sure if it was the graphics card, but they all worked fine.
I don’t know what I did wrong in the terminal section, or if I need to type in a different command line to get Fedora to load in, or if it actually was my graphics card. If I can’t find a solution, I might have to reinstall Fedora 44, which I don’t really want to do except as a last resort. Any help is appreciated, and I will try my best with my very limited understanding of the system to give any information you feel was left out or that I need to clarify.
There are several posts with problems involving the latest update on fedora 44 KDE and the same symptoms you are seeing. A quick search will show those.
I am not using KDE so was not affected but this seems critical for KDE users.
I doubt a reinstall would help since as soon as you did the first update the problem would reappear.
You might try using dnf and perform a rollback of the latest update using the updates-archive repo
I thought I did a thorough enough search on the forum, because all I got was posts from earlier versions that I didn’t think would resolve my current issue. I’ll try and be more thorough with my searches in the future.
So for a temporary solution use dnf to go back to a previous? And for a more permanent fix to my problem I would need to upgrade my graphics card to a more suitable version, am I getting that correct?
Thank you for answering my post. I really appreciate it. Once I get it working I’ll mark it as solved.
That may be possible, but it is also quite suitable for those who do not use graphics intensive apps such as gaming to simply use the nouveau driver and remain on wayland. For most desktop use the older nvidia gpus with nouveau and wayland meet the needs of many.
So I tried installing Plasma Workspace X11, but I’m not actually sure I did it right. So I went to the terminal and typed in sudo dnf install plasma-workspace-x11. It asked me if it wanted to install it I typed y. It installed, but I didn’t see any changes on the kernal version screen, and I still didn’t get past the black screen.
So I went to edit the current kernal version, and replaced the rhgb quiet with x11 to see if that would help in case I misunderstood where it went. I fear I’m not getting something. Can you tell me what I did wrong?
remove the x11 from where you put it in place of ‘rhgb quiet’
After installing that package you would need to select to use x11 on the login screen in order to tell the system which DE to use.
I don’t think step 2 above is required since the system worked properly until it was updated and is not related to either x11 or wayland but instead was apparently an update in plasma that triggered it.
What is required is that you identify the last update done that precipitated this problem then reverse it.
dnf history list | head -10 should show the last ten updates made.
sudo dnf install fedora-repos-archive to be able to use the updates-archive repo. (this step only needs to be performed once)
sudo dnf history undo TRANSACTION --enablerepo updates-archive where you replace the TRANSACTION with the id number of the indicated transaction that was listed and you intend to reverse that action. The transaction of interest would be the latest that shows as an upgrade.
Performing the 3 steps above should enable reversing the action that caused the black screen issue.
You can usually walk down the list shown in step 1 from the most recent backwards until a reboot works properly.
Thank you so much I feel like I’m starting to understand what to do now. However, I still ran into an error of sorts, or at least some kind of roadblock. I don’t know how to take and send a screenshot from a broken computer, so I’m going to write out what I did.
After logging back into the terminal, I typed in sudo dnf history list | head -10
It showed the dnf install plasma-workplace-x11 along with a bunch of other updates.
I then typed in sudo dnf install fedora-repos-archive. It was installed shortly afterwards. I then tried to type in sudo dnf history undo 34 --enablerepo updates-archive–allowrasing (I decided to uninstall the plasma-workspace-x11 first to make sure it works, before I go and undo the actual update(s)). A message then appeared that read- “No matching repositories for updates-archive–allowrasing. Add “–help” for more information about the arguments.” So I naturally, went to the add help option and this is what it gave me.
Usage:
Dnf5 [Global Options] history undo [Options] [Arguments]
Options:
–skip-unavailable
–ignore-extras
–ignore-installed
–store=STORED_TRANSACTION_PATH
–offline
Arguements:
transaction-id Transaction ID
I’m guessing these options are to either replace undo history or allowrasing, as William Yang said allowrasing cannot be used with undo. However I’m not sure which one is the best option for my current use. Again, thank you guys so much for being so patient with me, I really appreciate it.
That should have been sudo dnf history undo 34 --enablerepo updates-archive
I made an error in my instructions above as was pointed out by William above. (I fixed that error)
You also made a typo in not putting a space after the repo name and not using --allowerasing but shortened it to -allowerasing
Oh okay! I must of have misunderstood what you guys where saying, sorry about that. Thanks for informing me. I did put a - -allowrasing but in the post it must of smashed the two symbols together.
So it did undo the 34 plasma-workplace-x11 so that’s a win. But when I put:
sudo dnf history undo 33 - -enablerepo updates-archive.
It said:
Failed to resolve the transaction:
Cannot perform install action, no match for: then a string of letters and numbers (x5)
Extra package ‘steam-1.0.0.85-6.fc44.i686’ (with action ‘Remove’) which is not present in the stored transaction was pulled into the transaction.
It then says that I can either add the command line, skip-unavailable to skip unavailable packages, or ignore-extras to allow extra packages in the transaction. I’m not sure if I need to add one of those to the command line, and which is the one I need, or if I did something wrong again while typing the command again.
Okay so I have news, it worked! I had to retype it a couple of times into the command line until I got working right. The command I entered that worked:
sudo dnf history undo 33 - -enablerepo updates-archive - -skip-unavailable - -ignore-extras
It seemed to not only work, but I was finally able to get into the Fedora loading and login screen from uninstalling that one package. I will of course login in and out a couple of times to make sure it is working properly, and if I encounter the same issue I will follow the steps above.
It seems to have uninstalled my steam, but that’s fine for now. I am scared of updating it in case the same issue occurs again. I’ll look around more infomation to see if Fedora has fixed the issue. Thank you all so much- especially Jeff V -for your patience and support.
As I understand it, the issue was caused by an update of plasma. You should be able to update again by using a command such as sudo dnf upgrade --exclude plasma\* and things would work properly (cli only and not using discover).
The issue with that particular version of plasma is known and should be fixed very quickly, which then would allow full upgrades again.
Thank you! I typed the command line into the konsole as it’s currently working its magic. I do currently have a black screen, but my step-dad said it might take a few minutes after I reboot it for the updated kernel to work, hence the black screen. (Otherwise, if it persists, start at step one again )
I’m guessing after updating the kernel I’ll be able to reinstall steam again. Thanks again, much love to you.
Same Problem, but when I tried to run the command it dosnt allow it, becouse it would delete current kernel. I fixed this once by reinstalling but after the dnf Update it broke again
The same I have had with an NVIDIA gt1030 graphic card with the GP108 chip was lost after the new upgrade for Wayland , and now i have used the reserved internal gpu in the RYZEN 5 cpu and it has a little problem only to the secure boot , but be a little patiënt for an upgrade akmod from a new kernel upgrade it will repaired automaticly without problems.
)