[mythtv-users] scheduler settling with conflicts where they are resolvable

[Brian J. Murrell]

On 11-09-09 10:08 AM, Kevin Kuphal wrote:
> 
> I'm not sure what you're trying to say here.  You said yourself you resolved
> the conflict using the Upcoming Recording screens to tell the scheduler to
> record X on a particular tuner (essentially adding priority to it)

I'm saying there were two possible resolutions.  One was to shuffle the
channels that the desired recordings were being recorded on to get them
all to record on the multirec digital tuner.  My understanding is that
the scheduler won't try that hard to resolve conflicts.  Fair enough.

The second option to resolve conflicts was to assign the conflicting
recordings to other tuners that were going to be idle at the time of the
recording.  The scheduler didn't choose that option either (i.e. and
just left conflicts) and as a user, I have no ability to choose that
option.  But as I said, as a user, I should not have to tell the
scheduler to schedule a conflict on an idle tuner so I have no argument
with the lack of ability for me as a user to do that.

> Again, from your perspective.  From the scheduler's perspective, it is done.

How can the scheduler possibly be done when it has conflicts and idle
tuners!?!?  What's the point of the scheduler if it's not going to at
least try to utilize all of the resources it has to schedule with?

> It did the programmed number of passes,

If the scheduler really is so simplistic as to have a threshold of a
number of passes before it bails, why isn't that number configurable by
the user so that users that want to dedicate more resources (i.e. CPU
and time) to the scheduler can do so to avoid conflicts, even where they
are extremely simple to resolve -- like just utilizing another tuner.

> did the best it could do with the
> information given and the constraints on it and finished.

Let me decide the constraint then so that the scheduler can try harder.

> I sincerely doubt in your situation with the number of available tuners you
> appear have that a simple +1 preference on one program of the three that are
> typically recorded at the same time would have any problem.

But by that same measure, I should not have to +1 one of them.  There
are idle tuners!  How many passes does the scheduler actually do before
it even starts trying to use additional tuners?  Surely putting all of
the available tuners into the resource pool should be immediate and they
should be available for the first few passes even.

> I would suggest
> that you try it since you are looking for solutions and see what it really
> does instead of dismissing it.

Well, it seems I cannot try this test since something very hinkey is
going on.  I removed the override that I placed prior to force
everything onto the digital tuner and the new schedule is now using the
other tuners for the conflicts that were previously there.  I am
absolutely most definitely positive it was not before.

The scheduler has still left one conflict but that's because it's
arrangement of recordings on digital and analog tuners is less than
optimal and has exhausted the number of tuners.  I'd still argue that it
should try harder, and to that I'd argue that if there really is some
simple threshold value on the number of passes the scheduler makes and
if given an infinite value for that threshold the scheduler would
produce the most perfect fit schedule then that threshold should be user
configurable.  Let *me* decide how much resource I give to the scheduler.

At least tell me where in the code this threshold value is, please.

b.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 262 bytes
Desc: OpenPGP digital signature
Url : http://www.mythtv.org/pipermail/mythtv-users/attachments/20110909/8c0af175/attachment.bin