Upgrading Wireplumber in KDE effectively crashes Desktop

Help
,
1

I just did updates in Fedora 42 beta, I first updated dnf from 5.2.9 to 5.2.10.
Then I upgraded wireplumber from 0.5.7-1 to 0.5.8-1.
But each time the dnf transaction seemed to crash right after wireplumber was upgraded. The “dnf history info last” would show the Status as “Started”. I did that two times (#2 and #3 as a ‘reinstall’ ) then decided the “crash” was deterministic and not a random bug. I suspect wireplumber, since it has a sddm job, has a trigger that restarts sddm. So you can’t upgrade wireplumber from inside KDE, you have to log out and go to a virtual console or go whole hog and run a “dnf upgrade --offline” and “dnf offline reboot” to do the full blown upgrade at boot up.

Note also that just before all this I updated from dnf 5.2.9 to 5.2.10. So this might be a new picky thing that dnf 5.2.10 does where it insists on killing and restarting SDDM because it has a wireplumber job and absolutely has to have the latest and greatest wireplumber right now. :roll_eyes:
I’ve updated wireplumber/pipewire before from inside KDE and it’s never down this “kick out to SDDM” until now using dnf 5.2.10
Is this true?

BTW, these are the three dnf history “infos” for the two “crashes” to Status: “Started” and then the last one from the console which ends up Status: “Ok”

root:~
 dnf history info last-1
Transaction ID : 559
Begin time     : 2025-02-09 22:50:03
Begin rpmdb    : e931593638a813f13465024ccd7ae8ee1cfd299b4f86d1431fd547722438eb68
End time       : 2025-02-09 22:50:06
End rpmdb      : 3f59c0bccbde95d8748fde39ae3766f051e667510a58c25286662d17b8e39c8f
User           : 1000 Mark S <user>
Status         : Ok
Releasever     : 42
Description    : dnf reinstall wireplumber*
Comment        : 
Packages altered:
  Action    Package                                Reason     Repository
  Reinstall wireplumber-0:0.5.8-1.fc42.x86_64      Group      fedora
  Reinstall wireplumber-libs-0:0.5.8-1.fc42.x86_64 Dependency fedora
  Replaced  wireplumber-0:0.5.8-1.fc42.x86_64      Group      @System
  Replaced  wireplumber-libs-0:0.5.8-1.fc42.x86_64 Dependency @System

[20250209-180921]
# dnf history info last-2
Transaction ID : 558
Begin time     : 2025-02-09 22:49:23
Begin rpmdb    : e1bf27ba74cdd2397e07d114e47d0f892b2bb0e6d358c0c9caf2ba18e1b99a44
End time       : 1970-01-01 00:00:00
End rpmdb      : 
User           : 1000 Mark S <user>
Status         : Started
Releasever     : 42
Description    : dnf reinstall wireplumber*
Comment        : 
Packages altered:
  Action    Package                                Reason     Repository
  Reinstall wireplumber-0:0.5.8-1.fc42.x86_64      Group      fedora
  Reinstall wireplumber-libs-0:0.5.8-1.fc42.x86_64 Dependency fedora
  Replaced  wireplumber-0:0.5.8-1.fc42.x86_64      Group      @System
  Replaced  wireplumber-libs-0:0.5.8-1.fc42.x86_64 Dependency @System

[20250209-180927]
# dnf history info last-3
Transaction ID : 557
Begin time     : 2025-02-09 22:47:31
Begin rpmdb    : 4a5293591ba7c20ddb3d776cc4170290812574455eec3d2827a60127da306a77
End time       : 1970-01-01 00:00:00
End rpmdb      : 
User           : 1000 Mark S <user>
Status         : Started
Releasever     : 42
Description    : dnf upgrade wireplumber* -y
Comment        : 
Packages altered:
  Action   Package                                Reason     Repository
  Upgrade  wireplumber-0:0.5.8-1.fc42.x86_64      Group      fedora
  Upgrade  wireplumber-libs-0:0.5.8-1.fc42.x86_64 Dependency fedora
  Replaced wireplumber-0:0.5.7-2.fc42.x86_64      Group      @System
  Replaced wireplumber-libs-0:0.5.7-2.fc42.x86_64 Dependency @System
2

Hi - if you suspect this is mostly related to a dnf issue, then I’d suggest also making sure you are posting in the Fedora Discussion forums: Ask Fedora - Fedora Discussion - and it might be a worthwhile troubleshooting step to try the same operation with dnf4 to see if it is, in fact, tied to that version somehow.

Have you checked your system journal to see if there’s a specific indication there of what’s terminating your session?

3

The “dnf4 reinstall wireplumber” kicks me out of KDE just the same as dnf5 does. I get kicked to SDDM and have to login again.

There’s is this dnf.log (that would be dnf4) error about “input/output” error
related to the ‘reinstall wireplumber’

2025-02-10T13:56:17-0500 DEBUG ---> Package wireplumber.x86_64 0.5.8-1.fc42 will be reinstalled
2025-02-10T13:56:17-0500 DEBUG --> Finished dependency resolution
2025-02-10T13:56:17-0500 DDEBUG timer: depsolve: 115 ms
2025-02-10T13:56:17-0500 INFO Dependencies resolved.
2025-02-10T13:56:17-0500 INFO ======================================================================================================================================
==============================
 Package                                  Architecture                        Version                                     Repository                           Size
====================================================================================================================================================================
Reinstalling:
 wireplumber                              x86_64                              0.5.8-1.fc42                                fedora                              111 k

Transaction Summary
====================================================================================================================================================================

2025-02-10T13:56:17-0500 INFO Total download size: 111 k
2025-02-10T13:56:17-0500 INFO Installed size: 373 k
2025-02-10T13:56:20-0500 INFO Downloading Packages:
2025-02-10T13:56:21-0500 INFO --------------------------------------------------------------------------------------------------------------------------------------------------------------------
2025-02-10T13:56:21-0500 INFO Total                                                                                                                               163 kB/s | 111 kB     00:00     
2025-02-10T13:56:21-0500 DEBUG Using rpmkeys executable at /usr/bin/rpmkeys to verify signatures
2025-02-10T13:56:21-0500 INFO Running transaction check
2025-02-10T13:56:21-0500 INFO Transaction check succeeded.
2025-02-10T13:56:21-0500 INFO Running transaction test
2025-02-10T13:56:22-0500 INFO Transaction test succeeded.
2025-02-10T13:56:22-0500 DDEBUG timer: transaction test: 349 ms
2025-02-10T13:56:22-0500 INFO Running transaction
2025-02-10T13:56:22-0500 DEBUG RPMDB altered outside of DNF.
2025-02-10T13:56:22-0500 DDEBUG RPM transaction start.
2025-02-10T13:56:24-0500 CRITICAL OSError: [Errno 5] Input/output error
2025-02-10T13:56:24-0500 CRITICAL Traceback (most recent call last):
  File "/usr/lib/python3.13/site-packages/dnf/yum/rpmtrans.py", line 308, in callback
    self._script_start(key)
    ~~~~~~~~~~~~~~~~~~^^^^^
  File "/usr/lib/python3.13/site-packages/dnf/yum/rpmtrans.py", line 432, in _script_start
    display.progress(pkg, action, 100, 100, complete, total)
    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.13/site-packages/dnf/cli/output.py", line 1999, in progress
    self._out_progress(ti_done, ti_total, ts_done, ts_total,
    ~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                       percent, action_str, pkgname, wid1)
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.13/site-packages/dnf/cli/output.py", line 2026, in _out_progress
    print(" ")
    ~~~~~^^^^^
OSError: [Errno 5] Input/output error

but in the “dnf.rpm.log” I see:


2025-02-10T13:56:00-0500 INFO --- logging initialized ---
2025-02-10T13:56:22-0500 SUBDEBUG Reinstall: wireplumber-0.5.8-1.fc42.x86_64
2025-02-10T13:56:22-0500 SUBDEBUG Reinstalled: wireplumber-0.5.8-1.fc42.x86_64
2025-02-10T13:56:24-0500 INFO warning: %transfiletriggerin(systemd-257.2-17.fc42.x86_64) scriptlet failed, signal 1

2025-02-10T13:56:24-0500 ERROR Error in <unknown> scriptlet in rpm package wireplumber

For reference, the scriptlets in wireplumber are:

rpm -q --scripts wireplumber

preuninstall scriptlet (using /bin/sh):

 
if [ $1 -eq 0 ] && [ -x "/usr/lib/systemd/systemd-update-helper" ]; then 
    # Package removal, not upgrade 
    /usr/lib/systemd/systemd-update-helper remove-user-units wireplumber.service || : 
fi
posttrans scriptlet (using /bin/sh):

 
if [ $1 -eq 1 ] && [ -x "/usr/lib/systemd/systemd-update-helper" ]; then 
    # Initial installation 
    /usr/lib/systemd/systemd-update-helper install-user-units wireplumber.service || : 
fi
4

I’d suggest also making sure you are posting in the Fedora Discussion forums: Ask Fedora - Fedora Discussion

I’ve made a post in Ask Fedora that points back here

since there’s lots of info here I don’t want to replicate.

5

Since the DNF logs mention an input/output error, it would be worth looking in the system journal - using the journalctl command - to see if there is more information located there.

For example, if the attempted upgrade and crash happened at 3:30 PM today, you could use sudo journalctl --since="2025-02-10 15:29:00" --until="2025-02-10 15:31:00" to look at the surrounding couple of minutes and see what additional crashes, errors or warnings are being reported there.

Perhaps something is actually going wrong with your disk, or perhaps that’s appearing to be an I/O error because of something else crashing and interfering - looking at the core system logs should hopefully help narrow that down.

6

No, there’s nothing wrong with the disk