i wonder why this massive fail with akonadi and all the resulting blows haven’t been recognized in advance. that should not happen on a stable branch even if it’s rolling release.
6 Likes
After update I have lock screen freezes on Plasma with AMDGPU. When waking up after some time of inactivity the lock screen is not responsive. I can input my password, which is not showing up on screen, and can login after 10-20 seconds. I downgraded to 5.17 to see if it is kernel related.
Update completed without errors. I had to create the logfile for akonadi, but that worked fine. I had two issues. One seemed to resolve itself but the other one I had to work around.
First tracker-extract-3 slammed my CPU hard. I let it run and it finally stopped, but I think it was rebuilding something. I am running Plasma, but probably have some gtk stuff installed, so not sure what that was.
Latte-dock won’t start normally. It will start from a terminal without errors and will start with alt+f2 - latte-dock. However with auto-start or from the menu it just slams the cpu like tracker was doing.
I cannot reproduce your description using amdgpu and kde plasma 5.24.5, all amd system
The update itself went fine for me, but today I found that trying to launch VMs in VirtualBox while running the new 5.18.3 kernel results in the application hanging with a progress bar for “Starting virtual machine.” I end up doing a hard power reset because the process for VBoxSVC won’t terminate. Removing and reinstalling all the VirtualBox packages through Pamac did not help. Switching back to the 5.15.44 kernel does resolve the issue. I don’t have time to investigate further right now.
Edited to add: To be clear, Manjaro is my host system. I see other comments on the thread about VirtualBox, but it looks like they are talking about Manjaro as a guest system.
1 Like
The beep might be a result of the systemd upgrade, see this post in the testing branch:
I also have a problem with Pulseaudio 16.
When starting playback using Lollypop with PA16, my system hangs. Downgraded PA and all required libs to resolve this issue.
pacman -U pulseaudio-15.0-4-x86_64.pkg.tar.zst libpulse-15.0-4-x86_64.pkg.tar.zst pulseaudio-bluetooth-15.0-4-x86_64.pkg.tar.zst pulseaudio-equalizer-15.0-4-x86_64.pkg.tar.zst pulseaudio-jack-15.0-4-x86_64.pkg.tar.zst pulseaudio-lirc-15.0-4-x86_64.pkg.tar.zst pulseaudio-rtp-15.0-4-x86_64.pkg.tar.zst pulseaudio-zeroconf-15.0-4-x86_64.pkg.tar.zst
I’m using a sink to output audio to two displays using displayport via 6800xt (amdgpu).
after installing kernel 5.18 at reboot or shutdown I get the message:
“fsck clean…
Broadcasted message from machjine@user date…
Going to reboot or shutdown”
is it normal?
Same BEEP problem as others here have posted.
I tried the “blacklist pcspkr” solution from the testing thread, but it didn’t work.
The closing/shutdown message is different after update. It never used to have “Broadcast message from…”
I even tried to find an answer with MELD…nothing.
Thanks, unfortunately (blacklist pcspkr) did not have an effect.
It’s odd but the beep now also happens with my Nitrux Live USB. Im pretty sure it did not do that yesterday. Is it even possible?
Ill test with other live USB systems…
@Hanzel @worralorrasurfa @ColiZei
blacklist pcspkr does work! but the instructions in the testing thread are incorrect.
If you run echo “blacklist pcspkr” | sudo tee /etc/modprobe.d/nobeep.conf you’ll then need to edit the file & remove the " speechmarks. Then the beep is disabled.
Still stranger, this also fixed my Nitrux Live USB. (which uses systemd 219) ~ I cannot fathom how this manjaro setting can affect a completely separate live USB system.
3 Likes
Same error on F2FS with 5.15.46.
After fixing the issues with the missing file ib_logfile0 I found following errors many times in $HOME/.local/share/akonadi/db_data/mysql.err:
[ERROR] Incorrect definition of table mysql.column_stats: expected column 'hist_type' at position 9 to have type enum('SINGLE_PREC_HB','DOUBLE_PREC_HB','JSON_HB'), found type enum('SINGLE_PREC_HB','DOUBLE_PREC_HB').
[ERROR] Incorrect definition of table mysql.column_stats: expected column 'histogram' at position 10 to have type longblob, found type varbinary(255).
I executed following command to upgrade the database used by akonadi to actual mariadb version:
mariadb-upgrade --defaults-file=~/.local/share/akonadi/mysql.conf --socket=/run/user/$(id -u)/akonadi/mysql.socket
After restart of akonadi and mysqld the errors are not longer logged.
According Bug 454948 the db upgrade is not relevant for akonadi problems, but for me it avoided logging of unnecessary database errors.
1 Like
Same issue here. kernel 5.15.46-1 and VirtualBox 6.1.34 r150636.
100% CPU use, no keyboard / mouse response. The VM restarts after a few minutes.
2 Likes
if you read through pacman.log and have mariadb installed, you’ll notice the following TODO:
# systemctl restart mariadb.service && mariadb-upgrade -u root -p
With the addition of 5.18 and 5.19-rc1, have the recommended kernels updated as well? The app under settings manager still lists 5.15, 5.10, 5.4, and 4.19 as the recommended kernels (currently running 5.15 myself).
The recent update made my machine unbelievably LOUD. It used to be whisper quiet under moderate load, now fans are running full throttle in the same situation. The noise is unbearable. I’d be grateful for any suggestions on how to fix this. My machine has an X570 AORUS ELITE (rev. 1.0) with an Ryzen 7 (Pinnacle Ridge).
The only problem I’ve seen so far is that some Perl modules I installed from CPAN (I think, using compiled C code though I haven’t dug into it yet) come up with “undefined symbol” errors.
The fix seems to be to re-install them. There’s probably a way to identify them without waiting for Perl scripts to fail, but I don’t know what that would be.
Probably something like: pacman -Qoq /usr/share/perl5
amd_pstate error
To confirm: ignoring the error worked here, too, at least on kernel 5.18.3. Back on amd_pstate now. Far, far too lazy to try other kernels right now.
Seeing 590Mhz on cores again is a bit of a relief.
Hmmm… I suspect not. That returns a long list of packages, at least some of which were updated today. And others which were installed from AUR and haven’t been touched for some time (but still work).
I can’t see it helping with modules installed from CPAN either. But thanks for the suggestion.