[mythtv-users] My scheduler problems are back again

[Stephen Worthington]

On Thu, 5 Dec 2013 21:09:16 +1030, you wrote:

>
>
>> -----Original Message-----
>> From: mythtv-users-bounces at mythtv.org [mailto:mythtv-users-
>> bounces at mythtv.org] On Behalf Of Stephen Worthington
>> Sent: Thursday, 5 December 2013 3:51 PM
>> To: mythtv-users at mythtv.org
>> Subject: [mythtv-users] My scheduler problems are back again
>> 
>....
>> 
>> So, does anyone have any suggestions about how I might tweak the settings
>> to get the scheduler to assign the tuners better?
>> _______________________________________________
>
>Stephen, feel free to ignore me because I am far from knowledgeable about
>all this. But here goes anyway...
>
>With reference to your database.txt file and in particular the DVB-T inputs:
>
>1) Are you sure your cardid's and displayname's line up correctly with your
>physical tuners? To explain - I have 4 physical tuners each capable of two
>channels (8 total). But when I first set up my system the cardid's were not
>assigned neatly with cardid's 1+2 to the first physical tuner, 3+4 to the
>second etc. Instead physical tuner 1 got cardid's 1+5, physical tuner 2 got
>cardid's 2+6, etc. In a later rebuild (as I said I need a lot of learning) I
>tried to get smart and forced physical tuners1/2/3/4 to initially get
>cardid's 1/3/5/7 respectively but then mythtv assigned cardid's 8/9/10/11
>respectively instead of the hoped for 2/4/6/8 (which would have been neat
>and sequential like yours is). I also couldn't find a way of definitively
>mapping physical tuner to cardid but for me with a small number of tuners it
>didn't take much trouble to test and sort it out. I imagine it would be much
>harder to sort through 15 DVB-T channels - if your schedorder priority
>numbers actually put the same priority across different physical tuners that
>might precisely fit the symptoms you are seeing?
>
>Maybe to test - if you tune cardid 1 to a multiplex and keep it tuned in,
>are cardid's 2-5 restricted to channels on that multiplex as you would
>expect and prevented from tuning to a different multiplex?
>
>2) I see cardid's 1-5 all got schedorder 1, cardid's 6-10 all got schedorder
>5 and cardid's 11-15 all got schedorder 9. But I would have expected
>cardid's 1-5 to get schedorder 1, cardid's 6-10 to get schedorder 6 and
>cardid's 11-15 to get schedorder 10 (mostly based on how the schedorder
>defaulted on my system). Unfortunately this is more of an idle observation
>as I couldn't see it having any real impact on the behaviour you are
>currently noticing, or any real impact on anything at all.
>
>3) I see that your schedorder is the same order as your livetvorder. I think
>the recommendation is to have livetvorder opposite to schedorder to prevent
>(minimise) live tv stuffing up recordings but for the life of me I can't
>find the appropriate wiki reference. Maybe someone in the know can verify.
>
>Hope the above helps in some way.

I am one of those people who fiddle with my database.  So the reason
all the numbers line up is that I made them do it by changing things
manually after I added the new multirec tuners to increase them up to
the new maximum of five per physical tuner.  I find that it really
helps to have the numbering intuitive for a human when trying to debug
things, so I make my database work like that as much as possible.  And
due to the way the EPG works here in New Zealand, there is no way to
automatically add a channel anyway - if you let the automatic
processes happen, you wind up with a database that takes a lot of
manual configuration to make it work again, so I usually do channel
changes manually anyway, as it seems to take less work doing it that
way.

So, yes, all my cardids and cardinputids and displaynames do match,
because I made them match.  Without doing that, it would be quite
difficult to debug a problem like the one I am having, as I would all
the time be having to look up tables to figure out which tuner is
which.  Is it possible that I have made a mistake somewhere in the
setup?  Yes, of course it is.  But I have checked the capturecard,
cardinput and dtv_multiplex tables thoroughly a number of times, and I
have not found any problems.  And all the multirec tuners worked when
I tested them individually with LiveTV after doing the changes.

I did try your suggestion of seeing if I could have a DVB-T tuner
recording, then use LiveTV and switch to another of the multirec
tuners on the same physical tuner.  But I ran up against a LiveTV bug
or a DVBLOOPBACK bug - the last time I used LiveTV, it was tuned to
one of the channels from my pay satellite service that tunes on one of
the DVBLOOPBACK tuners.  And the first physical DVBLOOPBACK tuner was
recording from a channel on the same multiplex as the channel that
LiveTV was tuned to, using its first multirec tuner.  What happened
was that LiveTV attempted to tune the second multirec tuner on that
DVBLOOPBACK tuner, and got an error and stopped again.  So, due to the
way that LiveTV works, there is currently no way for me to change the
LiveTV channel to something else as LiveTV will not do that until it
has tuned its initial channel.  So I can not use LiveTV until the
current recording stops.  And another recording is going to start on
the same tuner a few minutes after that one stops, so if I do not
catch it in the gap and change the LiveTV channel, I will have the
same problem again.  I do wish there was a way to change the LiveTV
channel before it starts, but I have never found anything in the
database where it looked like it was the current LiveTV channel
setting, so I am stuck until the recordings finish.

The reason I have my livetvorder values the same as my schedorder
values is that I want LiveTV to try to use the same tuner as a
recording in progress (and vice versus) when the channel is on the
same multiplex as the physical tuner is tuned to.  That way, I should
not get LiveTV taking up a physical tuner and preventing recordings
from happening due to two physical tuners being tuned to the same
multiplex.  Or that was the thought anyway - I have no idea if it
actually works like that.