[mythtv-users] slow mythfilldatabase

Gary Buhrmaster gary.buhrmaster at gmail.com
Mon Jan 11 17:12:44 UTC 2021

On Mon, Jan 11, 2021 at 4:39 PM David Watkins <watkinshome at gmail.com> wrote:

>> I'd check to see if new listings are coming in. With the grabbers, expect
>> around 18 days of EPG.
> ... but beware, that doesn't necessarily mean you get 18 days of useful data.
> I was grabbing the full 18 days in the UK until I noticed that after 14 days most
> if not all of the programme titles were "To be Announced", which I imagine was giving mythfilldatabase a deal of unnecessary work.

Schedules Direct obtains data from Gracenote,
which for the US tends to have at some reasonable
data out to 17-18 days, but Gracenote depends
on country specific local sources outside the US,
so the data quality and duration can vary.  That
said, even for US guide data, sometimes all
one gets is a generic episode placeholder since
(for example) the network has not yet announced
the synopsis of an specific episode to be shown
more than a couple of weeks out even if it
believes it is going to show an episode of a
series in the timeslot (and for things like sports
programming, since even the teams playing may
not be known until a few days in advance, the
title is often generic).

> which I imagine was giving mythfilldatabase a deal of unnecessary work.

It uses some resources to do the processing, but
if the datetime/details are the same as previously
no database updates happen, so for the most
common use case (i.e. you aren't watching the
update like a hawk to time the job, impatient for
it to complete(*)), it tends not to be a huge resource
savings, but it also provides no value to have
data in the database that is not providing useful
scheduling information, so one should clearly
evaluate their use case.

(*) Typically, you pay close attention only when
you do the initial setup, and, of course, that
means all the data is new/different and that
is a lot of database updates, which can take
a lot of resources.  There are numerous reports
to this list of someone being surprised at
how long it takes to run their first/initial
mythfilldatabase and being concerned about
how long it takes, but shown that the next
updates tend to go a lot quicker for many
grabbers due to either the grabber caching
data itself, or the reduced database updates
for subsequent runs.

More information about the mythtv-users mailing list