DNF scriplet errors on F36: "Transport endpoint is not connected"

Hello,

Since my update from F34 → F36, I see a lot of failing scriplets in the DNF update history. They all have the following form:

Failed to start transient service unit: Connection reset by peer
Failed to set unit properties on PACKAGE.service: Transport endpoint is not connected

Until now, it affected the following package updates:
tracker
gnome-remote-desktop

What do these errors mean? Are they harmful (is the updated package broken), or is it simply a byproduct of those services not being in use at the time of the update? I have read in other forums that it might be an indicator of file system crashes and disk failure, but I don’t observe any other symptoms, and SMART also looks fine.

Thanks!

How did you update?
Did you follow this instructions:
Upgrading to a new release of Fedora :: Fedora Docs , If yes which one?

Yes, I performed the update according to the instructions using the DNF System Update plugin.

I made sure that my system was fully up to date before starting the update. After booting into F36, I performed the following post-update tasks:

Clean up retired packages (removed some packages)
Clean up old packages (nothing was found)

I performed the update on a KDE workstation spin (the “vanilla” one, not Kinoite, as it wasn’t released when I made my initial installation)

Thanks!

Here is a concrete instance of the error:

   3 Failed to start transient service unit: Connection reset by peer
   4 Failed to start transient service unit: Connection reset by peer
   5 Failed to start transient service unit: Connection reset by peer
   6 Failed to start transient service unit: Connection reset by peer
   7 Failed to set unit properties on tracker-miner-rss-3.service: Transport endpoint is not connected
   8 Failed to set unit properties on tracker-writeback-3.service: Transport endpoint is not connected
   9 Failed to set unit properties on tracker-miner-fs-3.service: Transport endpoint is not connected
  10 Failed to start transient service unit: Connection reset by peer
  11 Failed to set unit properties on tracker-extract-3.service: Transport endpoint is not connected
  12 Failed to set unit properties on tracker-miner-fs-control-3.service: Transport endpoint is not connected
  13 Failed to start transient service unit: Connection reset by peer
  14 Failed to set unit properties on tracker-xdg-portal-3.service: Transport endpoint is not connected

Given that the error references systemd service files, I have run systemctl list-units and systemctl list-unit-files. However, no services were found with the names the errors reference.

However, the service files referenced exist under the following path: /usr/lib/systemd/user/. Something seems to prevent systemd from finding these services, thus causing these errors during the update.

This is caused by a known SELinux policy issue that prevents systemd from restarting some services: 2100528 – %systemd_user_postun_with_restart spews errors and doesn't work

It should be harmless, it just causes some programs that should be restarted by an update to not be restarted. So logging out or rebooting will work around that problem.