[mythtv-users] Nearly 100% install of MYTHYV...

[Jerry McBride]

On Sunday 15 January 2006 21:00, Nick wrote:
> On 16/01/06, Jerry McBride <mcbrides9 at comcast.net> wrote:
> > Hi Nick,
> >
> > On Sunday 15 January 2006 19:55, Nick wrote:
> > > When I've seem MythTV failing to start properly using init scripts
> > > I've found that sometimes there is a /.mythtv folder with an
> > > incorrectly setup mysql.txt file. During init, root's home directory
> > > can be / in addition to /root, both of which can contain a .mythtv
> > > folder containing the DB details. Fixing up the contents of the
> > > /.mythtv/mysql.txt file fixed the DB connection details which had been
> > > stopping the mythbackend from starting.
> >
> > After running updatedb/slocate I found exactly 3 mysql.txt files. I
> > eliminated all but one and... it's still the same situation.
>
> Which one did you keep? I'd have cautiously kept all of them
> (identical copies) before removing them if the problem persisted.
>

I kept the one under /etc/mythtv. From strace output, this one gets referenced 
without problems.

> > Thanks for the tip though.
> >
> > > Starting the process when the system had fully booted (when the user's
> > > environment was going to be /root or /home/$MYTHBACKEND_USER) would be
> > > fine because the mysql.txt files for these users were correct.
> >
> > This makes sense to me save one thing... Even after the computer is fully
> > booted, try to execute the mythbackend script from a command line
> > fails... But executing mythbackend as a program succeeds.
> >
> > It's not a real show stopper by far, just one of those things I can't
> > fix...
>
> (probably nothing) Any chance that the extra "--" in the startup
> command (between my >> <<) is causing a problem?:
>
> --background >> -- << --verbose ${MYTH_VERBOSE} \
>

That ugly bit of bits is a separator for the start-and-stop utility... It 
caught my eye too... 

> Finally, have you checked the mythbackend log for entries before the
> backend failed? I don't see any mention of this in the thread, but it
> might contains very useful information, especially as the script can
> be run with verbose logging.
>

Yes. The backend server stops before anything gets written to it...

Once again, Cheers.


Jerry