KDE Plasma: on Rolling, crashes constantly

Hello,

Requirements:

I have Searched the forum for my issue and found nothing related or helpful
I have checked the Resources category (Resources Index)
I have reviewed the Wiki for relevant information
I have read the Release Notes and Errata

OpenMandriva Lx version:

OpenMandriva Lx release 26.02 (ROME) Rolling for x86_64

Desktop environment (KDE, LXQT…):
Operating System: OpenMandriva ROME 26.02
KDE Plasma Version: 6.5.5
KDE Frameworks Version: 6.22.0
Qt Version: 6.10.2
Kernel Version: 7.0.3-desktop-1omv2602 (64-bit)
Graphics Platform: X11
Processors: 16 × AMD Ryzen 7 9700X 8-Core Processor
Memory: 32 GiB of RAM (30.5 GiB usable)
Graphics Processor 1: AMD Ryzen 7 9700X 8-Core Processor
Graphics Processor 2: NVIDIA GeForce GTX 1660
Manufacturer: Gigabyte Technology Co., Ltd.
Product Name: B650 AORUS ELITE AX ICE
System Version: Default string-WCP

Description of the issue (screenshots if relevant):
I"ve been running Rock for 9 months or so on this same hardware – kernel 6.14
Never had an issue running KDE Plasma. Absolutely solid.
Maybe this is an X11 problem?
I finally upgraded to Rolling today.
Upgrade failed and had to do a complete install.
Anyways, now, KDE Plasma keeps on crashing.
Anyone else seeing this?
Is there a way to downgrade to earlier working version of KDE or anything?
Any help is much appreciated.

Things I’ve Tried

  1. I have integrated graphics - was running on those ports, but now changed to NVIdia – it crashes less but still crashes.
  2. turned off compositor – still crashes

Here’s what I see in logs:

Failed to register with host portal QDBusError("org.freedesktop.portal.Error.Failed", "Could > Aug 15 19:26:16 raddesk plasmashell[2133]: Failed to register with host portal QDBusError("org.freedesktop.portal.Error.Failed", "Cou> Aug 15 19:26:24 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:27:33 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:27:33 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:12 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:12 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:12 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:12 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:12 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:13 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:14 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:14 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 19:37:14 raddesk plasmashell[2133]: Exposed with no visual parent. Window positioning broken. Aug 15 22:14:36 raddesk plasmashell[2133]: The X11 connection broke (error 1). Did the X11 server die? Aug 15 22:14:36 raddesk ksecretd[1790]: The X11 connection broke (error 1). Did the X11 server die? Aug 15 22:14:36 raddesk systemd[1793]: Stopping plasma-plasmashell.service - KDE Plasma Workspace... Aug 15 22:14:36 raddesk plasmashell[2133]: qrc:/qt/qml/plasma/applet/org/kde/plasma/notifications/global/Globals.qml:132: TypeError: > Aug 15 22:14:36 raddesk plasmashell[2133]: qrc:/qt/qml/plasma/applet/org/kde/plasma/notifications/global/Globals.qml:262: TypeError: > Aug 15 22:14:36 raddesk plasmashell[2133]: ICE default IO error handler doing an exit(), pid = 2133, errno = 32 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 17 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 16 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 14 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 13 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 12 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 11 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 10 destroyed before end of thread 0x559a426017a0 Aug 15 22:14:36 raddesk plasmashell[2133]: QThreadStorage: entry 9 destroyed before end of thread 0x559a426017a0

Relevant informations (hardware involved, software version, logs or output…):

Obviously something went wrong somewhere in that install. Let’s see if we can get this working.

  1. Did you do the checksum on the ISO after you downloaded it?
  2. How did you write the ISO to the thumbdrive?

Thanks for your quick reply.
I downloaded the ISO and then used Balena Etcher to write it to the USB 3.0 flash drive.
However, I didn’t do a checksum on the original download.
I will do this as part of my standard practice in the future.
The download was on the original system where I did the fresh install so unfortunately I can’t go and check it.

I’m open to any suggestions. Should I download again and do a fresh install to fix this issue?

Did you install graphics drivers? I think the om-welcome app has an easy way to do nvidia, but I don’t know which driver version or kernel is needed for the 1660.

1 Like

I didn’t install graphics drivers. I think I was being a bit naive because I didn’t install graphics drivers seperately with OM Rock & everything worked great for all those months – never saw this issue.
I will take some time and check out the om-welcome app and how to install drivers.
Thank you for the suggestion. i will let you know how this turns out.

I used the om-welcome to install the nvidia drivers and when I did, I rebooted and then got two black screens with just a cursor.
I couldn’t do anything so I rebooted and switched from X11 to Wayland and I’m on that now, writing this up.

I am not averse to Wayland necessarily – if my desktop still performs the same.
In the past, the problem I had was that when I ran MS Teams I couldn’t share my screen properly — need it for work.

I guess I’ll run like this for a bit and see what happens.

Myself, I would do a fresh download of the ISO and do it again. Sometimes, things go wrong in an install and you can do a re-install a whole lot quicker than you can diagnose the problem and fix it, if ever. It may still not work properly, but you can rule that out.

2 Likes

downloaded from
https://sourceforge.net/projects/openmandriva/files/latest/download

openmandriva-6.0-plasma6-x11.x86_64.iso

Where is the hash / checksum that I should match?

I checked it locally with sha256sum and get:

05acf723de074768512af81a8280a79ecfa8b0979a611700d8fb3cd3f7a1388c

Perhaps it would have been better to ask questions at that step, the first time something goes wrong is the time to ask.

Be sure your ROME system is up to date by using the System Update utility. Beyond that I would think the first thing to try would be using Wayland, also try with one screen only. A guess would be to try Wayland with and without nvidia driver. Get that to work and go from there. If you really need to reinstall:

You access the checksums by clicking on the i with the circle around it. Perhaps consider using dd from terminal to copy install image to USB just in case. Balena Etcher should work but just in case revert to old tried and true methods.

If I am not mistaken the X11 ROME isos might already have Wayland available, you can change this at the sddm login screen in the lower left corner. If Wayland is not available you can install it with sudo dnf in task-plasma6-wayland or maybe sudo dnf in plasma-workspace-wayland. Last I did this task-plasma6-wayland worked just fine.

If you are going to try ugrading from 6.0 to ROME as I remember it this takes more than changing repos and simply running sudo dnf clean all ; sudo dnf dsync --allowerasing. I would need to test this to be sure and I don’t really have time to do that.

Downloaded the target one – which ended up being Rock.
Didn’t even know. Built a new install flash drive.
Installed.
Immediately switched the repos to Rolling.
Ran sudo dnf dsync
Fails!
Gets to the systemd install and goes blacks screen.
it is not possible to go from Rock to Rolling.

I’m not moving to Rolling / Rome.
I’m now back on Rock and will stay there indefinitely.
Wayland is bad.
The newer versions of KDE Plasma fail on X11
It’s all so bad. I’m very sorry that it is, but it is.
I have my stable system at kernel 6.14.
I don’t look forward to the Linux future. It’s very bad unfortunately.

1 Like

Why on earth would someone install Rock and immediately try upgrading to ROME? If you believe you want ROME install that.

As I tried to say upgrading from Rock to ROME is unfortunately not straight forward and simple. I suspect this is the problem you saw but can not say for sure with out evidence like terminal output or logs.

If we had evidence of the actual problem there is possibly a fairly easy fix. But we need evidence.

Edit: I am chasing this problem because I thought that problem with systemd was fixed but if it isn’t we need to get devs involved and see if it can be fixed. If you just want to stay on Rock, the safe choice admittedly, that is OK, I can understand that.

sudo dnf clean all ; sudo dnf dsync --allowerasing

Rock to Rome changes a lot of things. It is only safe to do that change from a TTY. Your best bet is to just install from a Rome iso.

1 Like

When I went to the Rome download area and clicked it took me to a Rock download iso
Couldn’t even tell til I booted the flash drive and started installing
It was unfortunate mistake
And I was just trying to let you all know that if you do that then immmedialtely go to rolling the install just black screens at system d

Some Notes

I got stuck in a bad loop
1 could only get Rome to run on Wayland with my graphics
But on Wayland I couldn’t run my backup software Deja Dups so I couldn’t get my backup of balena etcher
Downloading balena etcher wouldn’t run either
Finally I switched my hdmi cables to my AMD integrated graphics ports and got back to X11 enough to get the new iso onto USB

I also had updated my old i3 laptop to Rome and after that the WiFi continually failed
Even tho same laptop ran rock for 9 months no problems

I’m assuming these are related to the large number of kernel 7.x changers and Wayland changes

I had numerous software failures on Wayland which made it difficult to get back

Thankfully I’m back on Rock and will restore from backups

BTW - none of this is meant as a rant against the work at OM
OM is honestly top distro out there
I’ve run Ubuntu, Arch, Manjaro, Debian, Fedora and even tried Omarchy (blech ) but OM is the best to me
I’m sure OM team will get the challenges of 7.x worked out in future
Thanks for your help

I don’t have time to do this but I am working on a test Rock 2 ROME upgrade. That will be in it’s own forum thread when I have results to report.

I will either do this upgrade from Console Mode or a tty as suggested, and that is because the systemd upgrade crashes the graphic desktop. Obviously new users would not know to do this from Console Mode or tty unless they were searching this forum before they did the upgrade which most folks would not do. So maybe I’ll end up adding to Rock release notes also.

1 Like

FWIW I just did this test and successfully upgraded a Rock 6.0 system to ROME. However I had inside knowledge that there was a circumstance that requires a few simple extra steps. Above I had said there might be a problem with systemd that is incorrect, that problem was fixed. The problem is with sddm and is easy to work around if the user knows beforehand what to do.

1 Like

With regard to the WiFi issue did you check which power profile you were on in Rome. I had the same with an I5 installation and it was power saver turning the clocks down so far that the ring was going unstable. I ran cpupower as root, changed profile to ondemand and set a minimum frequency of 2600 rather than 800, had no further issues since.

1 Like

@bren1127

Welcome