I run a program called logcheck which regularly scans the logfiles looking for “interesting things”. One of these interesting things (to logcheck, not to me at first) was dbus complaining about mythtv. The logs look something like this:
Nov 17 05:08:18 elmo dbus-daemon[1621]: dbus[1621]: [system] Rejected send message, 2 matched rules; type="method_call", sender=":1.236" (uid=999 pid=3669 comm="/usr/bin/mythfilldatabase --verbose general --logl") interface="org.freedesktop.NetworkManager" member="GetDevices" error name="(unset)" requested_reply="0" destination="org.freedesktop.NetworkManager" (uid=0 pid=1930 comm="/usr/sbin/NetworkManager ")
What this is is the dbus daemon rejecting a message being sent from mythfilldatabase that wants to go to NetworkManager. Looking at NetworkManager’s dbus configuration, its doing all the right things.
So it’s a simple matter of just fixing up the dbus configuration and you’re done. Then I hit the first snag; the dbus documentation is a little sad. I remember seeing a show where they MRI scanned some artifacts to get to the writing and then used a maker-bot to recreate the hidden writing. They then painstakingly translated it using a team of experts to work out the English translation. Working out the dbus configurations is not quite that difficult, but it’s close. The easiest way is to copy something and see if it works, which is what I did.
I created a file /etc/dbus-1/system.d/mythtv.conf which permitted the myth user access to the GetDevices method. It looks like:
Reloaded dbus and all went quiet.
Related articles
- Dbus | C API (Part 1) (benzidwael.wordpress.com)
