Comment 4 for bug 220604

Doug (dscoular) wrote :

Hi,
I think this happens when the following cron job runs:

*/10 * * * * root [ -x /etc/init.d/mythtv-status ] && /etc/init.d/mythtv-status reload > /dev/null

Since it runs as root it fails to find ~/.mythtv/config.xml to find the backend.
It then attempts to use UPnP to find the backend. However, it fails to do this
and gives this rather misleading message about failing to load the MythTV
Perl API.

The reason you see it when you ssh it is that ssh is displaying the /etc/motd
which the cron job updates every 10 minutes. When you run it by hand it
runs under your userid which, generally has a ~/.mythtv/config.xml.

As a temporary workaround you could try the following:

sudo cp ~/.mythtv/config.xml /root/.mythtv

Within 10 minutes the cron job will run, update /etc/motd, and the error will
disappear.

The real bug is why the UPnP doesn't find the backend and that it mis-reports
the real issue.

Cheers,

Doug