[mythtv-users] Mythfilldatabase taking FOREVER

Larry Kennedy lunchtimelarry at gmail.com
Wed Jun 8 01:52:18 UTC 2022


On Sun, Jun 5, 2022 at 2:06 PM Gary Buhrmaster <gary.buhrmaster at gmail.com>
wrote:

> On Sun, Jun 5, 2022 at 3:06 AM Gary Buhrmaster
> <gary.buhrmaster at gmail.com> wrote:
>
> > There are numerous checks to determine if the
> > existing program is *really* the same (with all the
> > details).  Sometimes it is, sometimes it is not
> > ("generic" programs such as news often get
> > updates as the time gets closer to the "now"
> > value).
>
> And, as I recall, the generic (daily) news programs
> can sometimes be marked as repeats, and
> sometimes not, due to the logic that mythfilldatabase
> and the upstream scheduling source use.
>
> > And mythfilldatabase does some additional
> > adjustments of the data that can (next run) result
> > in detections of changes.  Some of which was
> > improved with recent (supported) releases of
> > MythTV.
>
> At the end of the run, there is a line of the form:
>       Updated programs: nnnnn Unchanged programs: nnnnnn
> which indicates how many programs were updated.
>
> If you run the same mythfilldatabase load immediately
> again (which implies the same source guide data) you
> will likely see a small (but persistent) number of updates
> due to how MythTV manages the program data and
> detection of changes (which require updates).  For
> a daily run, ~10% changes for Schedules Direct is
> not atypical as generic placeholder show data get
> replaced with actual episode data for the particular
> showings and new data for an additional day.
>
> Running a recent mythfilldatabase and a recent
> version of the grabber, which together support
> (mostly) accurate new (previouslyshown) marking,
> I tended to use --no-mark-repeats to prevent
> mythfilldatabase from "improving" the program
> data (you can also modify the database setting
> NewEpisodeWindow to accomplish something
> equivalent, both mostly only matter because
> Schedules Direct actually often has more than
> 14 days of program data, and using/changing
> either may cause different issues depending
> on the details (and --no-mark-repeats had
> a specific bug when used in older versions)).
>
> Lastly, sometimes database tuning may be
> appropriate.  From time to time there are
> discussions on the MythTV email lists, or
> forum, regarding database tuning,  Running
> mysqltuner.pl can point out specific cases
> which may be causing issues (just be aware
> that not everything mentioned by that script
> is actually going to be an issue for MythTV).
>

OK, Gary, I bit the bullet and upgraded to Ubuntu 20 LTS and Mythtv 31. I'm
sure that will help long term, but it sure created some additional drama :)

I found some old tuning parameters in mysql.cnf that I think were
problematic (thanks Bill from another post of yours I stumbled across).  I
removed those and it seems to help.  MFDB logging shows that
inserts/deletes are now taking 1 second or less, as seen here:

2022-06-07 21:47:44.317605 I  Removing existing program:
2022-06-13T14:00:00Z - 2022-06-13T15:00:00Z I57391.json.schedulesdirect.org
Little People, Big World
2022-06-07 21:47:45.123297 I  Inserting new program    :
2022-06-13T14:00:00Z - 2022-06-13T15:00:00Z I57391.json.schedulesdirect.org
Little People, Big World
2022-06-07 21:47:47.038876 I  Removing existing program:
2022-06-13T15:00:00Z - 2022-06-13T16:00:00Z I57391.json.schedulesdirect.org
Little People, Big World
2022-06-07 21:47:47.927278 I  Inserting new program    :
2022-06-13T15:00:00Z - 2022-06-13T16:00:00Z I57391.json.schedulesdirect.org
Little People, Big World
2022-06-07 21:47:49.853728 I  Removing existing program:
2022-06-13T18:00:00Z - 2022-06-13T19:00:00Z I57391.json.schedulesdirect.org
90 Day Diaries
2022-06-07 21:47:50.308645 I  Inserting new program    :
2022-06-13T18:00:00Z - 2022-06-13T19:00:00Z I57391.json.schedulesdirect.org
90 Day Diaries

Still, this is going to take longer than one would think necessary.
Especially, given that this is not the first run following the upgrade.

I have a few questions about why MFDB is slow:

1.  Does the number of channels in the Sqlite database matter?  I have
1,000 channels of which only 90 are marked as "selected."  Should I/can I
delete the unused channels?
2.  Does the number of channels in the mythtv database matter?  I have
1,300 channels, of which 100 or so are marked as "visible."  Should I/can I
delete the unused channels?
3.  I noticed that the XMLTV download to /tmp had over 1 million lines, per
"cat /tmp/xxx | wc".  That seems large.

Which of these is possibly slowing me down?  I assume #1, so I tried to
delete "unselected" channels from the sqlite database, but they got
reloaded later.  I can't recall what command I ran that caused the reload,
or if this was automatic and unavoidable.  Thoughts?


Larry

_______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://lists.mythtv.org/mailman/listinfo/mythtv-users
> http://wiki.mythtv.org/Mailing_List_etiquette
> MythTV Forums: https://forum.mythtv.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mythtv.org/pipermail/mythtv-users/attachments/20220607/bfb18c95/attachment.htm>


More information about the mythtv-users mailing list