In the debugging process, I upgraded Neon to 24.04, but there’s no change really.
The journal wentions something like startplasma-x11 core dumped, and startplasma-way core dumped. (I try to login sometimes with X11 and sometimes with Wayland).
To me, the process might appear quite complex. I guess I know how to install a specific version of a package or software. But I’m not sure how to roll back an entire upgrade. Also I worry about like a partial rollback where some packages would be older and some packages would be newer, and the system would be somewhere between the two I guess. I might worry about restoring the system to for example the up to date state, or maybe a previous state.
Also, thinking we’ve restored the system to a previous state. We want to update the system and stay current with the updates. I don’t know.
Let me know anyway Do you want to know anything about my system? Do you want me to issue some command and show you the output? I have access to TTY only more or less.
By the way, I tried a few things like startx and commands to start plasma, but I guess I don’t know how to do these things the right way. Things didn’t start with me.
Probably useful to know what happened just before the core dumped.
The core dump itself, might be useful to Plasma developers, but that would be if the problem lies within plasma itself, instead of some kind of configuration error.
First: from your post I understand you upgraded only DE "I upgraded Neon to 24.04.
And after you wrote: “how to roll back an entire upgrade”
Second: what package manager are you using?
Hmm so you were running an old Plasma implementation as there have been no Plasma or related KDE/Qt updates in the 22.04 base since the 24.04 upgrade in October 2024, so the recent update you had was a normal Ubuntu system one.
I can’t say what broke for you, but I wonder if some Ubuntu package that neon normally replaces or overrides may have been updated and slipped by (doesn’t seem likley) ? Or maybe a kernel security patch or driver update broke something?
Have you tried booting to a previous kernel at the grub menu?
To maybe see more info, at the login screen switch to a tty and try starting Plasma manually:
ctrl-alt f2 (or 3, 4, etc)
then startplasma-wayland
or startplasma-x11
There may be more specific errors shown before you get to the dump.
Not possible, unless you have installed the OS to a btrfs file system and created file system snapshots.
Timeshift is a great GUI for automating this, as well as for rsync based snapshots on non-btrfs installs, though I don’t know if this type works for full system upgrades.
I’m on the go right now. I have this image here from a few days back. Might be helpful with telling a bit about what happens before the dumped core. Thanks for your reply : )
At the very beginning, I did a normal package update. And that introduced the error we’re working on solving.
After that, I went on trying to fix the error. And someone suggested I do a “sudo do-release-upgrade” which would upgrade the Ubuntu base of the KDE Neon system. KDE Neon is based on Ubuntu.
Neon is based on Ubuntu, so apt is the package manager.
The thing that introduced the error to the system was a normal package update, not a system upgrade. Not sure what might have happened.
I tried booting to a previous kernel, which is a Linux 5 kernel, but I got the same issue.
I think I tried the suggested commands (which should issue starting the desktop environment I guess), and got the same things in the journal, but I might try again and let you know. Please check the pictures uploaded in this form that have red text. The error to be shown after running the suggested commands should show as what’s in these pictures.
Would you want me to look at something else other than the journal and show it to you. Let we know. Thank you for your reply : )
Yes, but that is only needed when migrating to a new base Ubuntu version - like 22.04 > 24.04, which happened 5 months ago, and only is needed every two years. An install after that time would already be on 24.04. I assumed that you decided to do a release upgrade quite late.
Apologies for the confusion!
I can’t tell much of what is going on in your images, sorry.
Okay I mean, My laptop has had neon for many years now. Maybe 5 years.
I hadn’t done any updates for a while. Hard to estimate how long. Maybe a year or two. Then I did an update about two months ago. And then the system broke after rebooting from the update.
After that, I went on debugging. And thought I would do the upgrade from 22.04 to 24.04 . That’s all. It was someone’s suggestion actually.
In any case. Any ideas about commands to type and share the output of? : )
ubuntu backported a mesa update to jammy. that means everything in the neon jammy archive would need rebuilding against the new mesa. unfortunately jammy is now EOL for neon so that isn’t going to happen
bluetoothd: sap-server seems to have a problem. Seems related to this, but not to our problem.
PAM is unable to find a shared object pam_lastlog.so and seems to still be going ahead with loading it. That seems to be a dependency for gkr-pam (GNOME Keyring). Assuming you have set up plasma (or at least sddm from what I see) to use the GNOME Keyring, that might be relevant.
Lastly, startplasma-x11 itself tells the exact line number where the problem is, which is inside syscoca
Considering both your screenshots, it seems that both, X11 and wayland sessions have the same error from the sddm-helper right before the plasma crash.
Depending upon how much you need to keep your keyring, maybe you can try changing it and see what happens?
After the update (not the system upgrade) that caused the crash, I did a system upgrade from 22.04 to 24.04.
…so this suggests an issue with the graphics drivers? Should I switch graphics drivers from Nvidia or intel (with Nvidia’l prime-select) to Nouveau? What do you think?
…there’s the red line that has ASSERT in it. It should be connected to the issue. Both with X11 and Wayland.
For example, running startplasma-wayland ruturns a few lines including a line very similar to this line, and then exits.
Any thoughts?
In general, my system should be very similar to a clean Neon install. I didn’t modify my system much. I only install software that shouldn’t interfere much with the system. Things like SQL and Aparhe server and Node.JS and videogames and 3D modeling software and Unity3D (not the Ubuntu Unity), etc etc etc… …many many things. But I shouldn’t have modified the system at all if I remember.
The change that broke the system was the update, that’s all. It was an update before an upgrade. Any thoughts or ideas?
Is there anyway to do something like reinstall the system without having to wipe out the entire partition? A thought comes to mind: something like a factory reset kind of thing? Hopefully without deleting files and packages not connected to the core system.
Hi. When you talk about keyring, I’m not sure if you mean that I can try getting rid of the data, which would be the passwords and so on, or whether I would be changing the software that handles authentication.
I can reset my passwords and so on. That would be completely fine if that’s the problem. But changing the keyring software or so on, I’m not sure I want to do that, only to keep things as unchanged as possible. I only don’t want to have my system too different from a default Neon install or a clean Neon install. …are you sure the problem is with keyring software? Please let me know.
(edit: oh and by the way, I can login to tty just fine : ) )
Would you want to get any more information about my system? Anything to type in the terminal and show you the output of? : ) Thank you for your reply and thank you for your time : )
I would normally assume that a KDE based distro does not use GNOME Keyring by default, but I might be mistaken. I use EndeavourOS and it uses KWallet.
You might want to continue with a KDE Neon user regarding this (I am not).
Yes, I was referring to changing the software and potentially losing the keyring data as a consequence.