[mythtv-commits] Ticket #12106: EIT: slow tuning on a channel restarts activescan (mplex eit starvation)

[MythTV]

#12106: EIT: slow tuning on a channel restarts activescan (mplex eit starvation)
----------------------------------+-----------------------------
 Reporter:  angela.schmid@…       |          Owner:  dekarl
     Type:  Bug Report - General  |         Status:  new
 Priority:  minor                 |      Milestone:  unknown
Component:  MythTV - EIT          |        Version:  Unspecified
 Severity:  medium                |     Resolution:
 Keywords:  eit mplex starvation  |  Ticket locked:  0
----------------------------------+-----------------------------

Comment (by angela.schmid@…):

 Unfortunately my cable provider sends a crippled EPG for about two years,
 no cure in sight.

 On one specific mplex it sends a 7 days EPG:
 Table_ID: 78 (0x4e)  [= Event Information Table (EIT) - actual transport
 stream, present/following]
 Table_ID: 79 (0x4f)  [= Event Information Table (EIT) - other transport
 stream, present/following]
 Table_ID: 80 (0x50)  [= Event Information Table (EIT) - actual transport
 stream, schedule]
 Table_ID: 81 (0x51)  [= Event Information Table (EIT) - actual transport
 stream, schedule]
 Table_ID: 96 (0x60)  [= Event Information Table (EIT) - other transport
 stream, schedule]
 Table_ID: 97 (0x61)  [= Event Information Table (EIT) - other transport
 stream, schedule]

 However, subtitles (episodetitles) are only provided in
 Table_ID: 78 (0x4e)  [= Event Information Table (EIT) - actual transport
 stream, present/following]
 Table_ID: 79 (0x4f)  [= Event Information Table (EIT) - other transport
 stream, present/following]

 On all other mplex a one day EPG with subtitles is send
 Table_ID: 78 (0x4e)  [= Event Information Table (EIT) - actual transport
 stream, present/following]
 Table_ID: 79 (0x4f)  [= Event Information Table (EIT) - other transport
 stream, present/following]
 Table_ID: 80 (0x50)  [= Event Information Table (EIT) - actual transport
 stream, schedule]
 Table_ID: 96 (0x60)  [= Event Information Table (EIT) - other transport
 stream, schedule]

 Therefor I have to scan all mplex, to get a 7 days EPG and having
 subtitles for one day.
 Before I have been scanning this specific 7 days mplex only, by editing
 the query in eitscanner.cpp. Since I noticed that other mplex send one day
 subtitles, I started scanning all mplex. Thereafter I noticed the
 described behavior.

 https://code.mythtv.org/trac/ticket/11809, wouldn’t solve my problem. I
 left a message there.

 I moved two weeks ago to Ubuntu 14.04, upgraded the DVB drivers and think
 everything is faster and more stable. I haven’t noticed tuning issues,
 even after I set the signal and channel timeouts from 3/6 to 1/1 second.

 For me the issue doesn’t exist anymore. Still I am curious where in the
 source Myth decides to restart active EIT scanning and why.
 Still I see the following issues in the algorithm: After recording on a
 configured EIT card, active EIT scanning is restarted with a random
 channel, instead of restarting the killed channel. If EIT channel scanning
 is 5 minutes and you have 40+ mplex it could take a long time the killed
 channel is rescanned, maybe missing the latest changes. In my case, after
 a tuning issue, the next channel, due to random selection, has a very low
 chance to be selected (starvation).

--
Ticket URL: <https://code.mythtv.org/trac/ticket/12106#comment:3>
MythTV <http://www.mythtv.org>
MythTV Media Center