[mythtv-users] Cover art

[Michael T. Dean]

On 04/01/2009 06:17 PM, Bobby Gill wrote:
>> The "file" of which you speak is most likely, "tmdb.xml".  That error
>> occurs when there is absolutely no response from themoviedb.org (not
>> even an error code) and, therefore, is /no/ data to parse.
>>
>> The interesting thing is that the error occurs /very/ quickly--in spite
>> of the timeout (which, by default with the Perl module in use is 3
>> minutes--I'm guessing, but haven't read the code to verify, that tmdb.pl
>> is using the default).  So, how can you get /no/ response extremely
>> quickly with a 3-minute timeout?  It happens when there's a network
>> issues--such as name-resolution failure.
>>
>> Run the script a 2nd time (after the caching-only name server you're
>> running has actually found the IP address for themoviedb.org) and it
>> works perfectly.  You can generally run it immediately after you dismiss
>> the error dialog.  (If you don't run your own name server, complain to
>> your ISP.  If your ISP's name server isn't causing problems, check your
>> network/routing/proxies/...)
>>
>> Mike "Hasn't gotten around to making a patch for that, yet, but it's
>> easy to work around by re-running" Dean
> The problem is, as soon as I hot OK to get out of that error dialog in the
> FE, the FE freezes on the "retreiving data.. [looking for] moviename"
> translucent screen.

Well, that shouldn't happen, and without a backtrace ( 
http://www.mythtv.org/wiki/Debugging and 
http://www.mythtv.org/docs/mythtv-HOWTO-22.html#ss22.2 ), I have no idea 
what could be causing a lockup.

However, could you test that's the only issue you have by "priming" your 
name server just before a test.  I.e. run:

ping -c 1 api.themoviedb.org

in a terminal, then switch back to the frontend and do the lookup.  The 
TTL for api.themoviedb.org is 24hrs, so really, you'd only need to do 
that once a day (though fixing your name server/the script is the 
preferred approach :).

If it fails after priming your name server, you have some other network 
issue causing problems (which may also explain why it "locks up"--though 
I'm guessing it's just doing something else, which is likely failing, too).

Mike