[Stable Update] 2022-06-12 - Kernels, Systemd 251, GNOME 42.2, KDE Gear 20.04.2, NVIDIA, Mesa, Pulseaudio, Pipewire, LibreOffice

After this update the below message shows on tty at KDE GUI Shut Down (the word poweroff is replaced by reboot on Restart) when clicking the button in the KDE Application Launcher.

I did some research and I believe this is caused by the updated systemd issuing commands to poweroff and reboot without the --no-wall argument (a new requirement?). Is there a way to correct locally or would it need to be patched upstream?

I do not get the beep from the pc speaker that others are reporting, maybe because I am on a laptop.

^@^@^@
Broadcast message from xxx@xxx-pc (day date time):

The system is going down for poweroff NOW!

_

Kernel missing, please load the kernel first
same here

Cannot boot Manjaro anymore. I’m not linux expert so… Please help!
How do I restore kernels or I how can I restore from Timeshift?
Timeshift made a backup before upgrade.

I fixed it myself. I booted from Manjaro Live USB.
I run Timeshift. It found the last backup before upgrade!!! So I was able to restore it to previews fully working state.
Manjaro and Timeshift are great! So I boot up my system again without the update.

I had done the last update with pacman -Syuu.
This time I made the update with pamac-manager. Update worked ok.
I don’t know why it replaced kernel real time 5.17rt for 5.18rt. It didn’t upgrade 5.17 normal to 5.18.

I’ll try to manually install normal 5.18 kernel with pamac-manager.

Two more broken items:

After the update, ffmpeg4.4 came up as an orphan, and I don’t know why. I know Audacity needs this because without it an error message pops up at the start of Audcity. (This may not happen to everyone. I typically go beyond the norm with Audacity)

  • Easy fix, just reinstall ffmpeg4.4.

Gnome-Boxes is now unusable. Any virtual systems created before the update still work fine, but now it will not boot and create a new virtual system. I downloaded and tried 4 different Linux distros. No luck.

I have Manjaro Xfce on another desktop that never had Gnome Boxes installed, so, I installed it just to see what would happen. Same result on a fresh install…It will not boot/create a new virtual box.

after login my desktop stuck at splash screen. here is my journald log

I am waiting to update, but wow, so many broken itens…
I think I will wait a little more… ¯_ (ツ)_/¯

Any news about VirtualBox?
It’s the only thing that I am really concerned in this update

2 Likes

From what I’ve read, the VirtualBox instability issues are due to a kernel regression:

3 Likes

I am using 5.15 LTS and saw some problems with that too… Any way to fix, or it’s better wait?

1 Like

I am temporarily using an earlier version of the 5.15 kernel and VirtualBox kernel modules to avoid the instability issues. If you are not comfortable downgrading or holding packages, it might be better to wait until a patched kernel is released :).

Does anyone know which kernel versions have fixes for the regressions?

1 Like

this is not good ,
this is pacman -Syu or pacman -Syyu

Since you used CPAN you cannot ask your package manager, you have to look for yourself:

$ ls -l /usr/lib/perl5/5.34/*/
1 Like

And the same here, too. Affected VMs are Windows 10 and 11 guests.
Also tested both available 5.x LTS kernels with no difference. Didn’t test higher kernel versions, though.
I uninstalled an re-installed both the VM Extension Pack and the Guest Additions, but to no avail.
I even downgraded VirtualBox to the previous v6.1.32 with the same negative result, which is why I suspect either the kernel updates or the systemd update to cause this trouble.
Unfortunatly my VMs don’t run long enough to take a glimpse of their event logs to gain more insight from that side.

2 Likes

Nice one. Thanks!

After this update wine 7.10 stopped working properly with the MetaTrader4 program (timeframes changing stopped working and blank program window starts when Alert window appears).

Downgrading wine to a previous 7.7 version was a workaround solution solving the issue.

Downgrading kernel, respective VBox host modules and Virtualbox to the previous versions did the trick for me for now.

just to report (nothing critical for me, I’ve downgraded, all fine, will wait for the next kernel)
On Plasma (ext4, nvidia but no gaming, fairly simple setup),

linux518 fails (I can log in but I get ‘kwin is too unstable’ dialog), I’ve had to reboot into 5.17, remove 5.18.
I’ve looked into the journalctl, there is what it seems a kernel segfault around that time

"kernel: MainThread[1095103]: segfault at 0 ip 00007fdbabeec7b5 sp 00007fff6e7ab9c0 error 6 in libxul.so[7fdbaabae0
00+5a82000]"

followed by a core dump

systemd-coredump[2181]: Process 2164 (kwin_x11) of user 1000 dumped core.

to my eye this looks like libxul.so is where it crashes and called from firefox module - what’s firefox doing at this point?

…that repeats number of times, bunch of other errors (plugins missing, skipped) but I couldn’t make much sense of it.

The system has been running ok for a long time (year, more), this is the first issue of this sort after an update. i did mirror fasttrack & sudo pacman -Syyu - so fairly standard for me, nothing special this time.

I didn’t see anyone else reporting this - and that is strange, so I’ll just wait it out as this seems like a major kernel update.

And thanks guys for all the work you’re doing!

EDIT: wrong dump : ) - similarly, there are multiple core dumps for kwin_x11

As @LA-MJ points out, during upgrade, mariadb prompts one to upgrade the databases.
The prompt during updating, however, only concerns the general mariadb databases
(systemctl restart mariadb.service && mariadb-upgrade -u root -p)
which you may perhaps not even be using or running.

Akonadi runs its own and one can check whether a database upgrade is needed (it was in my case)
mariadb-upgrade --defaults-file=~/.local/share/akonadi/mysql.conf --socket=/run/user/$(id -u)/akonadi/mysql.socket --check-if-upgrade-is-needed

I guess, backing up ~/.local/share/akonadi/db_data/ib_logfile0 is a good precaution (as it seems to regularly disappear on first reboot after the update).
After updating and rebooting, I copied my backed up ib_logfile0, started akonadi, checked whether akonadi’s databases need an upgrade, performed the mariadb-upgrade, rebooted, and all seems fine.

1 Like

Update: This is limited to Electron Flatpaks. Same electron apps from the repos work fine, which is probably better, but still this is very odd. Seems to be related to a display driver of some sort, according to a two year old reddit thread.

Do you remember the steps you made?
I also noticed comments about upgrading the MariaDB (MySQL) database, but i struggled so much trying to do it that at the end i simply recreated the logfile. Most notably, the one thing that really blocked me in that upgrading process was that, on my end, the socket was missing, and Akonadi never managed to recreate it, until i went for the easy solution.

There is no chance to classified this VirtualBox issue was “Known issue” from this update?
That’s the only thing that I am waiting a solution to update my machine.

I saw another post and they said that upgrading the kernel to 5.18 kind of fix the problem.
This really work?

I had the same issue, but with my bt-headphones (Sony WH-1000XM4). I was able to fix it by replacing pulseaudio-modules-bt with pulseaudio-bluetooth as well, but I also had to change my /etc/pulseaudio/default.pa adding ‘load-module module-bluetooth-discover’