EndeavourOS disappeared from boot menu after a yay update

Hello! I’ve just went through a very stressful sequence of events:

  • I have a Thinkpad T480, set up for dual boot of EndeavourOS and Windows (I need it for some tasks)
  • I was doing some work on Windows today
  • After that, when I rebooted into EndeavourOS I was having some trouble connecting my bluetooth headphones - hasn’t been a problem before.
  • I decided to run a yay update to see if that fixes it
  • To my horror, when I rebooted after running yay, I was no longer able to log into EndeavourOS. On an attempt, an error would very briefly flash on the screen: /usr/bin/startplasma-wayland: symbol lookup error: /usr/lib/libKF6KIOGui.so.6: unidentified symbol: _ZN14KWaylandExtras18xdgAc [here’s where the photo I took cuts off - I stupidly didn’t bother to check if I got all of the message in the frame. It wasn’t much longer than this though, probably a few more characters]
  • I tried rebooting into the fallback and logging in there, but the result was the same
  • I rebooted once again and EndeavourOS was no longer in the boot menu, so I’m writing this from Windows.

I assume I’ll have to boot into the live USB and fix my install with chroot. Is that correct?

If so, sorry for a bit of a duplicate post, I just need some reassurance this is fixable. I’m currently not home and won’t be able to boot into live USB for a couple hours.

Yes, this would be the first option. Boot with live system and chroot into your installation. Then run a complete update. Afterwards you have to reinstall grub or systemd-boot depending on what you are using.
Maybe the error of SDDM gets fixed by this.

Ok, I’ve managed to chroot into my system, but I can’t reinstall systemd-boot. What am I missing here?

I don’t think you need to install bootctl. Maybe just try sudo reinstall-kernels while in arch-chroot.

Edit: Since there is a new kernel anyway maybe just run another update first and see in arch-chroot.

Yes, that’s correct. I reinstalled the kernel and it got EOS back to the boot menu.

I’m still getting that weird error when trying to log into plasma, though. I don’t really understand what is happening or why but is it safe to assume it’s an issue with plasma itself?

I can boot into shell just fine, so I could probably try installing a different DE and use that for a while.

What happens when you try to log in?

Edit: You could try installing lightdm and the enable and start the service.

I’m currently using ly as my display manager and when I try to log into plasma I get the following error: /usr/bin/startplasma-wayland: symbol lookup error: /usr/lib/libKF6KIOGui.so.6: undefined symbol: _ZN14KWaylandExtras18xdgActivationTokenEP7QWindowRK7QString (note: I was only able to take a blurry photo due to the short time the text was visible so I may have misread some characters from it)

It flashes on the screen very briefly, then I’m thrown back to login. The same happens with SDDM, but the screen only briefly turns black and I do get to see the error message. I am able to log into Weston.

Have you tried lightdm?

I’ve just tried it. It fails to start.

Did you enable and start the service?

systemctl enable lightdm.
systemctl start lightdm

Yes. Here’s what I get:

To better see what errors KDE is throwing, try this disabling lightdm and sddm. When your system reboots, you’ll be at a console login, After logging in, run this:

/usr/lib/plasma-dbus-run-session-if-needed /usr/bin/startplasma-wayland

This will at least remove sddm as the culprit. I’ve been running without a display manager, so I know this works (updated my KDE systems yesterday). If this does not work, then maybe try downgrading kio.

Trying to log in directly from the console results in the error from earlier:

So whatever is causing it is the real culprit.

It’s late now, so I’ll leave this for tomorrow and go to sleep. If nobody finds a solution to this, I’ll just treat this whole ordeal as a sign to finally learn how to set up and use hyprland.

Edit: Forgot about looking into downgrading kio, will try that later.

That - the downgrading - did, in fact, work. Thank you so much! I chose kio 6.4.1 kind of arbitrarily. Which version is the latest that is compatible?

Also, do you have any idea why would all of this happen in the first place?

Edit: I went back up to kio 6.18.0 and while I can log in now, some bits of Plasma are broken. KRunner will not start and application launcher displays: file:///usr/share/plasma/plasmoids/org.kde.plasma.kickoff/contents/ui/main.qml:21:1: Cannot load library /usr/lib/qt6/qml/org/kde/plasma/private/kicker/libkickerplugin.so: /usr/lib/libKF6Runner.so.6: undefined symbol: _ZN14KWaylandExtras18xdgActivationTokenEP7QWindowRK7QString

Hi,

I downgraded kio to 6.18. That works for me.

There is an update to kio this morning so maybe that fixes the issue?

Sadly it doesn’t. I assume the problem isn’t inherent to kio but some other part of Plasma that for some reason isn’t installed correctly on my system. For now I’ll keep using kio 6.18 and just accept that some things will be broken, until I figure out where to go from here.

Sometimes i question whether it’s an update or some change made specifically. Otherwise we’d all be having the same problem. I say this because i rarely make changes to my system and I don’t run into these problems. :person_shrugging:

As far as I can tell there could be a problem with Krunner go to system settings and in the search box type Krunner click plasma search>configure krunner and maybe disabling it or empty the history cache could help. Might also be a plugin you have downloaded/installed for Krunner

For me empty the cache did not work, I had to downgrade krunner to 6.18 too.