Created attachment 13570 [details]
list of recent package updates

I am now remoted into the cauldron machine and displaying claws-mail on mga8 laptop; this is running fine, so I'm now suspecting KDE/plasma issues.

Trying to access audio settings:
$ systemsettings5
systemsettings5: symbol lookup error: /lib64/libQt5QuickWidgets.so.5: undefined symbol: _ZTI22QAccessibleQuickWindow, version Qt_5_PRIVATE_API

Created attachment 13571 [details]
journalctl -b

Fired up claws-mail on affected system. An empty window appears (actually, contents of lower window) for 30 seconds, then CM opens as it should have. Click on inbox, displays with no delay. Select another message: quick. Another: 30 sec delay.  After this, CM is acting normally...  WTH?

Closed, re-opened; started normally...  ???

systemsettings5 still fails to start.

This is confusing...

I'm seeing lots of plasma errors in the journal; attached so you can see it asap while I try to dig through it...

Regarding the Payload SHA256 ALT digest: BAD message, try "urpmi --clean" and
then "urpmi --auto-update".

I'm getting a black screen for plasma in my m9 vb guest so am currently using
IceWM.

CC: (none) => davidwhodgins

$ urpmi --clean
root@pf   Tue Dec 13 18:28:21
~
$ urpmi --auto-update
    http://mirrors.kernel.org/mageia/distrib/cauldron/x86_64/: media/core/release/media_info/20221213-223523-synthesis.hdlist.cz
...retrieving failed: aria2 failed: exited with 3
problem reading synthesis file of medium "Core Release"
medium "Core Updates" is up-to-date
medium "Nonfree Release" is up-to-date
medium "Nonfree Updates" is up-to-date
medium "Tainted Release" is up-to-date
medium "Tainted Updates" is up-to-date
Packages are up to date

Tried a few times; always get the Core Release error.

Hope to comment shortly on journalctl output...

There's a lot in the journal. Start by changing the downloader to wget in
drakrpm-edit-media, Options, Global options. Then run urpmi --auto-update

Once that's done, for the message
dkms-autorebuild.sh[1200]:   Kernel devel files for 6.0.12-server-1.mga9 not installed.  Cannot install this module.

install ...
kernel-server-devel-latest
kernel-server-latest

Make sure the user(s) who will be running vb are in the vboxusers group
- run userdrake
- doubleclick on the userid
- switch to the groups tab and ensure vboxusers is selected

I recommend installing htop which has more options then top. See what's taking
a lot of time.

I suspect akonadi will take quite a while to index everything that's readable
by the user with that much space. Be sure to leave the system running long
enough for it's activity to be reduced to near zero.

reboot

I've created bug 31276 for the messages
systemd-udevd[549]: /usr/lib/udev/rules.d/69-libftdi.rules:2 Unknown group 'plugdev', ignoring

Where did /etc/udev/rules.d/80-evrouter.rules come from? It's not from
a Mageia package and has errors.

After the reboot, create a new log file with "journalctl -b --no-hostname".

$ drakrpm-edit-media
Ignore the following Glib::Object::Introspection & Gtk3 warnings

(drakrpm-editmedia:1391763): dbind-WARNING **: 23:12:50.703: Couldn't connect to accessibility bus: Failed to connect to socket /tmp/dbus-io4OoTYFPh: Connection refused
Subroutine Gtk3::main redefined at /usr/share/perl5/vendor_perl/Gtk3.pm line 539.
getting exclusive lock on urpmi
Use of uninitialized value in concatenation (.) or string at /usr/share/perl5/vendor_perl/urpm/cfg.pm line 237.
wrote config file [/etc/urpmi/urpmi.cfg]
wrote config file [/etc/urpmi/urpmi.cfg]
unlocking urpmi database
(In reply to Dave Hodgins from comment #7)
> There's a lot in the journal. Start by changing the downloader to wget in
> drakrpm-edit-media, Options, Global options. Then run urpmi --auto-update

Do the errors matter?
$ drakrpm-edit-media
Ignore the following Glib::Object::Introspection & Gtk3 warnings

(drakrpm-editmedia:1391763): dbind-WARNING **: 23:12:50.703: Couldn't connect to accessibility bus: Failed to connect to socket /tmp/dbus-io4OoTYFPh: Connection refused
Subroutine Gtk3::main redefined at /usr/share/perl5/vendor_perl/Gtk3.pm line 539.
getting exclusive lock on urpmi
Use of uninitialized value in concatenation (.) or string at /usr/share/perl5/vendor_perl/urpm/cfg.pm line 237.
wrote config file [/etc/urpmi/urpmi.cfg]
wrote config file [/etc/urpmi/urpmi.cfg]
unlocking urpmi database


> Once that's done, for the message
> dkms-autorebuild.sh[1200]:   Kernel devel files for 6.0.12-server-1.mga9 not
> installed.  Cannot install this module.
> 
> install ...
> kernel-server-devel-latest
OK

> kernel-server-latest
Always been there.

> Make sure the user(s) who will be running vb are in the vboxusers group
> - run userdrake
> - doubleclick on the userid
> - switch to the groups tab and ensure vboxusers is selected

VB has always been part of the install; but I never selected it since I have no need for it.  

> I recommend installing htop which has more options then top. See what's
> taking
> a lot of time.

Been an htop user for years.  It showed nothing obvious.  In my original comment, I mentioned 30, 60 and 90 seconds which implies timers which don't contribute to %CPU, etc.

> I suspect akonadi will take quite a while to index everything that's readable
> by the user with that much space. Be sure to leave the system running long
> enough for it's activity to be reduced to near zero.

That's an interesting comment...  according to https://userbase.kde.org/Akonadi, akonadi is used by  KMail, KOrganizer, KJots, and KAddressBook; none of which I use -- although I've tried many times to setup KOrganizer which always errors out.

> reboot
> 
> I've created bug 31276 for the messages
> systemd-udevd[549]: /usr/lib/udev/rules.d/69-libftdi.rules:2 Unknown group
> 'plugdev', ignoring
> 
> Where did /etc/udev/rules.d/80-evrouter.rules come from? It's not from
> a Mageia package and has errors.

No idea; I have never looked at udev rules, let alone messing with them. 
A search for evrouter leads to https://www.bedroomlan.org/projects/evrouter/ which I've never heard of; but since it's related to routing mouse events, I'd guess it's related to my Logitech MX Master 3 and likely "solaar"...
 
> After the reboot, create a new log file with "journalctl -b --no-hostname".

I get ...
$ drakrpm-edit-media
Ignore the following Glib::Object::Introspection & Gtk3 warnings
Subroutine Gtk3::main redefined at /usr/share/perl5/vendor_perl/Gtk3.pm line 539.
getting exclusive lock on urpmi
wrote config file [/etc/urpmi/urpmi.cfg]
wrote config file [/etc/urpmi/urpmi.cfg]
unlocking urpmi database

Please attach /etc/urpmi/urpmi.cfg from that system.

I take it you were able to install kernel-server-devel-latest despite the
errors.

solaar (which I use too) does not have any udev rules.

What's the date of the file /etc/udev/rules.d/80-evrouter.rules

Created attachment 13572 [details]
journal after changes and reboot

Things have gotten worse...  

Now, it boots to a black screen**; no plasmashell (it keeps trying) -- just one konsole and Mageia Welcome.

** you got this too apparently (comment 5)

The black screen reported as bug 31280

To use IceWM, at the login screen click on the down arrow beside the session,
and select it.

To access the menu in IceWM, right click on the desktop.

(In reply to Dave Hodgins from comment #9)
> I take it you were able to install kernel-server-devel-latest despite the
> errors.
Yup
 
> What's the date of the file /etc/udev/rules.d/80-evrouter.rules
-rw-r--r-- 1 root root 49 May 26  2022 80-evrouter.rules

Created attachment 13573 [details]
urpmi.cfg

It is perhaps rather obvious to mention it, but Cauldron is in a very turbulent state. The Beta ISOs are not good...
Despite updating this Cauldron system often enough, it has just swallowed an update of hundreds of pkgs. Despite which, it has never broken down even though some things (sound notably) have stopped working.

This is the real 'beware' territory warned of for Cauldron users. If big updates go wrong, it is often the case that re-trying some time later works.

I specifically tried those originally cited complaints (comment 0) that I could, but without 'focus follows mouse'; no emacs. This is under LxQt, but perhaps I should try under Plasma. I will only add additionally if that differs from:
* claws-mail: selecting another message takes 60 seconds
  - clicking menu (File) takes 90 seconds
  - having moved mouse across menu bar since clicking; File menu tries to open,
    but then tries to open another menu item -- this doesn't complete (partial
    menu list opens for a while). 
Claws-mail continues to work normally here.

* firefox: 100% CPU; won't switch tabs; totally unresponsive - killed.
Works normally. [Alas, recent versions have lost the ability to see BBC videos].

CC: (none) => lewyssmith

"turbulent" is the operative word, even for the current problems...  :)

At the moment, the main issue has become bug 31280 -- without plasmashell, this bug is moot.  Getting the correct Qt lib should take priority...

The server parts (ftp, ssh, postgreSQL and an AWS watcher script) are all running fine. Only the desktop is affected; and I am handling that by using my old laptop as the plasma desktop. Not nearly as easy with only one much smaller screen; but workable while I await a fix for bug 31280...

(In reply to Dave Hodgins from comment #7)

> 
> Where did /etc/udev/rules.d/80-evrouter.rules come from? It's not from
> a Mageia package and has errors.
> 

Google shows at least one instruction for creating this file:

"The event devices (/dev/input/event*) need to be readable by normal users. You need to create a udev rule to accomplish this task:
KERNEL=="event*", NAME="input/%k", GROUP="input"
/etc/udev/rules.d/80-evrouter.rules"

at "Getting the Logitech K350 multimedia keys to work properly in Linux"

https://www-user.tu-chemnitz.de/~klada/?site=projects&id=logitechkbd

Rolf

CC: (none) => rolfpedersen

Rolf is right...  I forgot about the K350 extra keys. Never got that to work; maybe I'll try again.

BTW, plasmashell is back with the latest updates.  Thanks!!  Now to see if the weird delays were impacted by 31280... after about 10 minutes, looking good. Fingers crossed.

Nope...  still having problems with long delays. Trying to find a pattern.  So far, symptoms are mostly inconsistent; except:
* starting claws-mail took a while before the windows were populated; I run 2 instances of CM -- both did that the first time after a reboot; but exit and re-open both instances of CM was quick.  3rd time was quick too... 

Operative words:  inconsistent symptoms

Something that may be related:  pulseaudio has been screwed up since first connecting the APC UPS -- not connected and won't be tried until everything is stable again.  Moments ago, I was running alsamixer and checking for anything obvious:
$ alsamixer
pfortin@pf   Wed Dec 14 18:39:11
~
# Looked OK; but seconds later, trying to restart it:
$ alsamixer
ALSA lib pulse.c:242:(pulse_connect) PulseAudio: Unable to connect: Connection refused

cannot open mixer: Connection refused
pfortin@pf   Wed Dec 14 18:39:16

Through several reboots, using "play" always hung; but this time, it just worked when "time"d:

$ time play /usr/share/sounds/Oxygen-Sys-Error-Printing.ogg
ALSA lib pulse.c:242:(pulse_connect) PulseAudio: Unable to connect: Connection refused


/usr/share/sounds/Oxygen-Sys-Error-Printing.ogg:

 File Size: 41.4k     Bit Rate: 160k
  Encoding: Vorbis        
  Channels: 2 @ 16-bit   
Samplerate: 48000Hz      
Replaygain: off         
  Duration: 00:00:02.07  

In:100%  00:00:02.07 [00:00:00.00] Out:99.3k [      |      ]        Clip:0 [NOTE]   
Done.

real    0m2.449s
user    0m0.010s
sys     0m0.073s
pfortin@pf   Wed Dec 14 18:38:23
~

Usually, it just hung at [NOTE] with the duration stuck at the audio clip's length, not [00:00:00.00] as above. I was not expecting to hear the audio clip; "time" was intended to see if it returned on its own...  Now audio plays; but opening System Settings - Audio right now returns a blank/empty panel, and alsamixer gets connection refused.

The reason I suspect pulseaudio is that the app delays may be due to an interaction with sound...

The sound/delay problems persist through reboots; though the only reboot not tried as I write this is a complete power down...  Will try that next.

Until the sound bug is fixed properly, uninstall all packages shown by
rpm -qa|grep -e pipewire -e wireplumber
using "rpm -e --nodeps $packagename". Then reboot.

Just did that and power down reboot -- now plasmashell won't start:

$ plasmashell --replace
kf.plasma.quick: Applet preload policy set to 1
qml: PlasmaExtras.ScrollArea is deprecated. Use PlasmaComponents3.ScrollView instead.
file:///usr/share/plasma/packages/org.kde.desktoptoolbox/contents/ui/ToolBoxContent.qml:238: TypeError: Cannot read property 'visible' of null
file:///usr/share/plasma/packages/org.kde.desktoptoolbox/contents/ui/ToolBoxContent.qml:237: TypeError: Cannot read property 'text' of null
qml: PlasmaExtras.ScrollArea is deprecated. Use PlasmaComponents3.ScrollView instead.
file:///usr/share/plasma/packages/org.kde.desktoptoolbox/contents/ui/ToolBoxContent.qml:238: TypeError: Cannot read property 'visible' of null
file:///usr/share/plasma/packages/org.kde.desktoptoolbox/contents/ui/ToolBoxContent.qml:237: TypeError: Cannot read property 'text' of null
"file:///usr/share/plasma/shells/org.kde.plasma.desktop/contents/views/Panel.qml" 
 "Error loading QML file.\n12: plugin cannot be loaded for module \"org.kde.taskmanager\": Cannot load library /usr/lib64/qt5/qml/org/kde/taskmanager/libtaskmanagerplugin.so: (libKPipeWire.so.5: cannot open shared object file: No such file or directory)\n"
Could not set containment property on rootObject

$ rpm -q -f /usr/lib64/libKPipeWireRecord.so.5.26.3
lib64kpipewirerecord5-5.26.3-1.mga9

When I tested removing all of the pipewire packages, I was using IceWM. I'd
reinstalled them to try to figure out another way to fix the issue, which
I've now proposed as a fix for bug 31275

Oops. bug 31235

I believe problems are fixed?

CC: (none) => fri

Must be; operating normally now...

Thanks

Status: NEW => RESOLVED
Resolution: (none) => FIXED