Bug 1411299

Description of problem:
I'm running Fedora in docker container and as a such it doesn't run systemd nor bus. If I try to install/upgrade/remove package which has systemd unit rpm trigger is invoked and runs systemctl daemon-reload. This isn't successful and prints "Failed to connect to bus: No such file or directory".

Version-Release number of selected component (if applicable):
systemd-231-3.fc26.x86_64

How reproducible:
Always

Steps to Reproduce:
1. docker run -ti fedora:rawhide bash
2. dnf upgrade --downloadonly -y dnf
3. cd /var/cache/dnf/rawhide-*/packages
4. rpm -Uvh *.rpm

Actual results:
In the output after dnf and libcomps I see:

Cleaning up / removing...
  26:dnf-yum-1.1.9-6.fc26             ################################# [ 58%]
  27:dnf-1.1.9-6.fc26                 ################################# [ 60%]
Failed to connect to bus: No such file or directory
  28:python3-dnf-1.1.9-6.fc26         ################################# [ 62%]
...
  44:librepo-1.7.18-4.fc26            ################################# [ 98%]
  45:libcomps-0.1.7-6.fc26            ################################# [100%]
Failed to connect to bus: No such file or directory

Expected results:
No messages: Failed to connect to bus: No such file or directory

Additional info:
My rawhide Image has sha256:3bcdeb6ee43bf3dfffd130c5fe6b81d47b04aed315638a85fbf26760962d733e. I'm not sure how to pull it by this id though.

I'm testing this in docker but I guess this can happen in mock, anaconda or other common environments where systemd is not running.

File triggers.systemd (http://pkgs.fedoraproject.org/cgit/rpms/systemd.git/tree/triggers.systemd, upstream: https://github.com/systemd/systemd/blob/master/src/core/triggers.systemd.in) seems to be the code which fails here. I guess it could check if /var/run/dbus/system_bus_socket exists and reload it only if it does.