[mythtv-users] backend always hangs on specific recording

[Stephen Worthington]

On Wed, 12 Jun 2013 14:06:33 -0700, you wrote:

>On Wed, Jun 12, 2013 at 4:36 AM, Stephen Worthington <
>stephen_agent at jsw.gen.nz> wrote:
>
>> On Tue, 11 Jun 2013 21:06:09 -0700, you wrote:
>>
>> >Tuesday afternoon at 1:30pm, mythtv is set to record 2 shows, and every
>> >week when I come home Tuseday evening, I find that the backend is stuck:
>> >the backend can't connect to it, and later scheduled recordings didn't
>> >happen.
>> >After killing and restarting the backend so that the frontend can connect
>> >to it again and get the list of recordings, I see one recording shows up
>> >with the "missing file" X next to it, while the other shows up as a 0
>> sized
>> >recording. If I try to play the 0-sized one, the backend seems to hang
>> >again.
>> >I can manually tune to the channels in question just fine, and I've
>> >recorded two things simultaneously from those two channels too. There's
>> >nothing in the log that indicates any kind of failure. Any idea why those
>> >same scheduled recording would always cause the backend to hang?
>>
>> What version of MythTV are you using?
>
>
>This is 0.25 (Mythbuntu, in case that matters).

OK, so if your problem is the same as mine, upgrading to 0.26 should
make mythbackend exit instead of lock up, and then you can have it
automatically restarted as Mythbuntu 12.04 does and recordings will be
(re)started as soon as mythbackend is running again.

> I used to get a similar problem
>> to that on, IIRC, 0.24.  I upgraded to 0.25 and then quickly to 0.26,
>> so I am not sure if 0.25 ever did it.  It was a deadlock situation,
>> and I tended to get it at the same time each week, but it took more
>> than two programs at once to trigger it for me.
>
>
>I regularly record 2 or 3 programs at once without any issues. The case I
>described above is the only one that consistently reproduces the problem.

My deadlocks were happening typically in more complex recording
situations than you are describing - I would have one or two
recordings already in progress and at least two new recordings
starting at the same time.  But like you, I also had many, many
occasions where I had similar numbers and types of recordings
occurring where the deadlock did not happen, including situations more
complex than where it occurred.  But when the deadlocks were
happening, if the recording schedule did not change next week, they
would happen again at the same time.  I have never been able to figure
out just what particular combination of conditions is needed to
trigger the problem.

>
>>  In 0.26, I can still
>> get a similar problem, but instead of a deadlock, mythbackend detects
>> the lockup and restarts itself.  On restart, it records OK.  This is
>> much better than having it stop recording, but I still miss a couple
>> of minutes of recordings due to the time mythbackend takes to restart.
>>
>> In my case, if I already had recordings running at the time of the
>> deadlock, those recordings would continue and end normally, but any
>> new recordings would never start after that.  If it is like my
>> problem, you may be able to prevent it from happening by getting one
>> or other of the recordings to start a minute or two earlier.