[mythtv] LiveTV bug
Bruce Markey
bjm at lvcm.com
Sat Mar 1 21:28:41 EST 2003
Matt Zimmerman wrote:
> On Sat, Mar 01, 2003 at 08:20:49PM -0500, Joseph A. Caputo wrote:
>
>
>>If the front end is killed (i.e., via a signal) while watching LiveTV,
>>then when the front end is restarted it won't allow you to watch LiveTV
>>again because "all inputs are being used for recording". There is no
>>in-progress recording in the list to delete, since it was LiveTV, so the
>>only option is to kill the backend and restart.
>
>
> Then don't do that.
But Doctor, it hurts! ;-)
> Seriously, if you are killing the frontend with a signal, why is it a
> problem for you to do the same thing to the backend?
It's simplest to think of starting and stopping the frontend
and backend at the same time but each is subject to dying,
being killed, a machine is rebooted or crashes, etc. Either
end should be well behaved even if the other is not. Currently,
the Live TV recording does successfully exist if the frontend
goes away ever since Isaac patched this last week.
The opposite should also work. If the backend is restarted,
the frontend should notice and reconnect without needing to
be restarted. I suspect that this will be fixed, or will be
easier to fix, after the multiple backend code is in place.
I assume the frontends will need to check to be sure they have
a connection before making requests to a backend.
I'd hate to think that we'd need to stop and start every
frontend and backend on every machine each time one
process dies.
-- bjm
More information about the mythtv-dev
mailing list