[mythtv-users] recording problems with recent 'fixes' build

[Udo van den Heuvel]

Udo van den Heuvel wrote:
> mythbackend takes ~100% of cpu and doesn't really respond anymore to 
> mythweb; also it doesn't record anymore.

> 2008-12-20 09:08:29.345 TVRec(4): ASK_RECORDING 4 29 0 0
> 2008-12-20 09:08:29.965 TVRec(2): ASK_RECORDING 2 29 0 0
> 2008-12-20 09:08:30.008 TVRec(6): ASK_RECORDING 6 29 0 0
> 2008-12-20 09:08:30.070 TVRec(3): ASK_RECORDING 3 29 0 0
> 2008-12-20 09:08:30.119 TVRec(7): ASK_RECORDING 7 29 0 0
> 2008-12-20 09:08:30.127 TVRec(5): ASK_RECORDING 5 29 0 0
> 2008-12-20 09:08:43.112 AutoExpire: CalcParams(): Max required Free 
> Space: 3.0 G
> B w/freq: 5 min
> 2008-12-20 09:12:00.357 TVRec(4): Changing from RecordingOnly to None
> 2008-12-20 09:12:00.371 Finished recording Z at ppSport: channel 3
> 2008-12-20 09:13:44.124 AutoExpire: CalcParams(): Max required Free 
> Space: 3.0 G
> B w/freq: 8 min
> 2008-12-20 09:17:00.095 TVRec(3): Changing from RecordingOnly to None
> 2008-12-20 09:17:00.110 Finished recording Herberg De Verandering 
> "Hijltje Vink"
> : channel 2
> 2008-12-20 09:21:44.273 AutoExpire: CalcParams(): Max required Free 
> Space: 3.0 G
> B w/freq: 15 min
> 2008-12-20 09:36:44.453 AutoExpire: CalcParams(): Max required Free 
> Space: 3.0 G
> B w/freq: 15 min
> 2008-12-20 09:51:44.511 AutoExpire: CalcParams(): Max required Free 
> Space: 3.0 G
> B w/freq: 15 min
> 2008-12-20 10:02:00.137 TVRec(2): Changing from RecordingOnly to None
> 
> (now it is 12:06 local time 24h format)
> 
> So this `fixes` build is buggy and really in need of fixes?

A cron job to reschedule at ~ 7:

2008-12-20 07:05:01.711 Using runtime prefix = /usr
2008-12-20 07:05:01.712 Using localhost value of localhost
2008-12-20 07:05:01.740 New DB connection, total: 1
2008-12-20 07:05:01.758 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 07:05:01.761 Closing DB connection named 'DBManager0'
2008-12-20 07:05:01.762 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 07:05:01.767 New DB connection, total: 2
2008-12-20 07:05:01.768 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 07:05:01.771 Current Schema Version: 1214
2008-12-20 07:05:01.927 Connecting to backend server: 127.0.0.1:6543 
(try 1 of 5)
2008-12-20 07:05:01.975 Using protocol version 40
2008-12-20 07:05:02.058 Connected to master for reschedule

(looks OK?)

And one at ~11 o'clock:

2008-12-20 11:05:02.090 Using runtime prefix = /usr
2008-12-20 11:05:02.092 Using localhost value of localhost
2008-12-20 11:05:02.150 New DB connection, total: 1
2008-12-20 11:05:02.204 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 11:05:02.205 Closing DB connection named 'DBManager0'
2008-12-20 11:05:02.209 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 11:05:02.213 New DB connection, total: 2
2008-12-20 11:05:02.217 Connected to database 'mythconverg' at host: 
localhost
2008-12-20 11:05:02.225 Current Schema Version: 1214
2008-12-20 11:05:02.284 Connecting to backend server: 127.0.0.1:6543 
(try 1 of 5)
2008-12-20 11:05:09.298 MythSocket(9a965c8:5): readStringList: Error, 
timeout (quick).
2008-12-20 11:05:09.298 Protocol version check failure. The response to 
MYTH_PROTO_VERSION was empty.
2008-12-20 11:05:09.298 Cannot connect to master for reschedule

By  then the backend had stopped recording although at 7 there was for 8 
hours worth of program info due to the custom searches I use to record 3 
channels 24/7.

What logging do I enable to find the cause?

Udo