TV Server won't stop recording! (1 Viewer)

[gibman]

Look in method
void StopRecord(RecordingDetail recording)
in scheduler.cs

U have one big try statement.
Thats ok - except that the finally block is deleting the schedule item from the collection object.
so if anything generates an exception in this method while trying to stop the record, then the record is also deleted from this list - even though we failed somewhere.

I simply moved the "_recordingsInProgressList.Remove(recording);" away from the finally block.

The _recordingsInProgressList object is being monitored by a timer thread - so it knows which items are to be stopped etc.
But not if they are removed

/Gibman

 

[Elf]

Well, so far so good - with the newest SVN installed I scheduled a recording which completed succesfully...

Interestingly, in the tswriter.log these lines appear at the beginning of the recording:

10-10-2007 22:22:17 Recorder:add private stream pid:25b
10-10-2007 22:22:17 Recorder:program stream mode
10-10-2007 22:22:17 Recorder:Start Recording:'I:\Documents and Settings\Admin\My Documents\My Recordings\Look North-BBC ONE\Look North-2007-10-10-22_25.mpg'
10-10-2007 22:22:20 CPesPacket::Fullbuffer 1208 1208
10-10-2007 22:22:25 CPesPacket::Fullbuffer 237 237
10-10-2007 22:22:27 CPesPacket::Fullbuffer 237 237

...Those "CPesPacket::Fullbuffer" errors keep occurring every few seconds until...

10-10-2007 22:40:10 CPesPacket::Fullbuffer 1864 1864
10-10-2007 22:40:12 CPesPacket::Fullbuffer 1864 1864
10-10-2007 22:40:14 Recorder:Stop Recording:'I:\Documents and Settings\Jason\My Documents\My Recordings\Look North-BBC ONE\Look North-2007-10-10-22_25.mpg'
10-10-2007 22:40:14 del m_pVideoAnalyzer
10-10-2007 22:40:14 analyzer: reset
10-10-2007 22:40:14 del m_pPmtGrabber
10-10-2007 22:40:14 del m_pRecorder
10-10-2007 22:40:14 del m_pTimeShifting
10-10-2007 22:40:14 del m_pTeletextGrabber
10-10-2007 22:40:14 del m_pCaGrabber
10-10-2007 22:40:14 del done...
10-10-2007 22:40:14 CMpTsFilter::Pause()
10-10-2007 22:40:14 CMpTsFilter::Stop()

perhaps those "fullbuffer" errors were the exception in the StopRecord() method that caused the recording to be deleted from the list?

I'm going to try some more scheduled recordings (if I can find any relatively short programmes starting in the next few minutes!) and will post more.

I guess the question then becomes "what are those fullbuffer errors"?



--------<edit>

Another success! This is looking promising!

The logs, however, now include a lot (in excess of 60-per-second) of odd messages like this:

10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 782 new: 1 days 00:21:36 460 prev: 1 days 00:21:36 459 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 1 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 782 new: 1 days 00:21:36 461 prev: 1 days 00:21:36 460 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 0 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 785 new: 1 days 00:21:36 464 prev: 1 days 00:21:36 461 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 3 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 789 new: 1 days 00:21:36 467 prev: 1 days 00:21:36 464 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 3 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 792 new: 1 days 00:21:36 471 prev: 1 days 00:21:36 467 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 3 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 795 new: 1 days 00:21:36 474 prev: 1 days 00:21:36 471 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 3 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 800 new: 1 days 00:21:36 479 prev: 1 days 00:21:36 474 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 5 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 828 new: 1 days 00:21:36 506 prev: 1 days 00:21:36 479 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 27 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 856 new: 1 days 00:21:36 534 prev: 1 days 00:21:36 506 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 27 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 858 new: 1 days 00:21:36 536 prev: 1 days 00:21:36 534 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 1 - pid:1a21
10-10-2007 23:25:42 hole: 0 days 00:00:00 0 hi: 0 days 00:16:27 890 new: 1 days 00:21:36 568 prev: 1 days 00:21:36 536 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 31 - pid:1a21

-------<edit again>

Another success - looks like that may have sorted it, gibman!

Can any of you other guys that were getting this confirm it seems OK with the latest SVN?



The logs have also settled down now, with the previous burst of data stopping with the following (something to do with timeshifting then?):

10-10-2007 23:26:33 hole: 0 days 00:00:00 0 hi: 0 days 00:17:18 121 new: 1 days 00:22:26 800 prev: 1 days 00:22:26 777 start: 1 days 00:05:08 678 - diff: 0 days 00:00:00 22 - pid:1a21
10-10-2007 23:26:33 Timeshifter:Stop timeshifting:'I:\Program Files\Team MediaPortal\MediaPortal TV Server\live1-0.ts.tsbuffer'
10-10-2007 23:26:33 Timeshifter:Reset

The third successful recording, for what it's worth, looked like this:

10-10-2007 23:37:07 Recorder:transport stream mode
10-10-2007 23:37:07 Recorder:Start Recording:'I:\Documents and Settings\Admin\My Documents\My Recordings\Taking a Line for a Walk-More 4\Taking a Line for a Walk-2007-10-10-23_40.ts'
10-10-2007 23:37:07 Pcr new start pcr :0 days 10:03:19 939
10-10-2007 23:37:07 Recorder: PMT version changed from -1 to 23
10-10-2007 23:37:07 Recorder: pmt:0x110 pcr:0x1ffe video:0x24e audio1:0x24f audio2:0x250 audio3:0 audio4:0x0 audio5:0x0 video:0x24e teletext:0x0 subtitle:0x0
10-10-2007 23:37:07 Recorder: first PMT change, ignore it!
11-10-2007 00:00:07 Recorder:Stop Recording:'I:\Documents and Settings\Admin\My Documents\My Recordings\Taking a Line for a Walk-More 4\Taking a Line for a Walk-2007-10-10-23_40.ts'
11-10-2007 00:00:07 del m_pVideoAnalyzer
11-10-2007 00:00:07 analyzer: reset
11-10-2007 00:00:07 del m_pPmtGrabber
11-10-2007 00:00:07 del m_pRecorder
11-10-2007 00:00:07 del m_pTimeShifting
11-10-2007 00:00:07 del m_pTeletextGrabber
11-10-2007 00:00:07 del m_pCaGrabber
11-10-2007 00:00:07 del done...
11-10-2007 00:00:07 CMpTsFilter::Pause()
11-10-2007 00:00:07 CMpTsFilter::Stop()

 

[fairchild]

Look in method
void StopRecord(RecordingDetail recording)
in scheduler.cs

U have one big try statement.
Thats ok - except that the finally block is deleting the schedule item from the collection object.
so if anything generates an exception in this method while trying to stop the record, then the record is also deleted from this list - even though we failed somewhere.

I simply moved the "_recordingsInProgressList.Remove(recording);" away from the finally block.

The _recordingsInProgressList object is being monitored by a timer thread - so it knows which items are to be stopped etc.
But not if they are removed

/Gibman

strange it just started working for me... changed absolutely nothing on my end, which SVN was this changed in?

 

[gibman]

Thanks for testing it all ...

Its in SVN since 08/10/2007

the tswriter logs is just excessive debug logging - it should be disabled in latest svn (according to tourette).
I am not sure about the severity of the "CPesPacket::Fullbuffer" entries.

/Gibman

 

[jimbolina]

I'm still getting a lot of the "hole: 0 days..." logs in my tswriter.log with yesterday's svn. The file is over 150MB after running it all night last night.

https://forum.team-mediaportal.com/tswriter_log_eating_my_hdd-t29411.html

 

[djm66]

I installed the latest TV3 SVN yesterday and everything is working great; recordings are starting and stopping as they should! Hooray!

You really miss TV Server when you go back to standard mediaportal; It's a great bit of software when it works.

I do have the same problem a Jimbolina though; a 50mb TSWRITER log file with millions of "hole.." lines, looks like Tourettes is onto this tonight though.

 

[djm66]

I spoke too soon!

TVServer did a 43 hour recording over the weekend for a 1 hour programme! Noooooooooooooo.

The problem is still here (using Rev16179 - 11 oct 2007) .. NOT FIXED

Some recordings start and stop without a problem, others (normally longer recordings never stop until hard drive space runs out).

I want to be more helpful and post logs but they don't appear to be that useful.. TSWRITER log file is so huge I can't really access it, it's got millions of lines saying "hole..." and is MASSIVE! The TV log file starts today (14 Oct)..

Not sure what to do... would really like to keep using TVserver.

I will install the latest TV3 SVN and do some tests - when it happens again, I'll post full logs for client and server.

Anyone else still having this problem?

 

[gibman]

Just post all the logs - tswriter.log is of no use.

Also explain the whole scenario for us.

We might have to do some additional logging in order to solve this.

/Gibman

 

[djm66]

Ok, thanks for looking into this.

It happened again this morning; I scheduled TVserver to record Quantum Leap this morning (07.55 to 08.55 on 15 Oct 2007) on ITV3. Powerscheduler woke up my computer and started recording at 07.55. However, it did not stop recordng at 08.55 as scheduled, it only stopped recording when I switched on my monitor and started investigating (at 10.05) i.e. this problem seems only to occur when powerscheduler wakes up my computer from suspend. My PC is awake enough to start recording, but not awake enough to stop!

When my PC is always on and I'm doing stuff in the background, recordings seem to start and stop ok.

Logs are posted, I'm using TVserver Rev 16213 14 Oct 2007 in single seat mode.

Thanks again!

 

[gibman]

Thanks for the logs.

I need 2 know what kind of setup u have.

I assume u have 1 multiseat client and 1 tv server.

Which PC's does hibernate/standby and resume ?

I cant understand how hibernating/resuming a multiseat client will affect the outcome on a tv server.

powerscheduler, is that the tve3 plugin ?

Myself, I am using 24-7 tvserver which never goes into hibernate, only my dumb clients are
Thus, I am not using powerscheduler on tvserver.

/Gibman