[confirm] Why oh Why :( Read() no open file (2 Viewers)

[mylle]

1Gbit wired network.

 

[miroslav22]

Thanks

At least that narrows it down a bit further. Very strange though whatever I do on my system I can't get it to happen. It ran quite happily on its own for hours with just the odd few late frames every few minutes.

I'll keep prying

 

[romuz]

Here logs like

23-01-2011 21:55:48.846 [710]Demux : Video to render 1.003 Sec
23-01-2011 21:55:49.135 [710]Demux : Video to render 0.964 Sec
23-01-2011 21:55:49.135 [710]Demux : Audio to render 0.628 Sec
23-01-2011 21:55:50.431 [710]Demux : Video to render 0.959 Sec
23-01-2011 21:55:51.562 [710]Demux : Video to render 0.899 Sec
23-01-2011 21:55:51.564 [710]Demux : Audio to render 0.603 Sec
23-01-2011 21:55:55.175 [710]Demux : Audio to render 0.590 Sec
23-01-2011 21:55:55.415 [710]Demux : Audio to render 0.578 Sec
23-01-2011 21:55:56.017 [710]Demux : Video to render 0.768 Sec
23-01-2011 21:55:56.020 [710]Demux : Audio to render 0.464 Sec
23-01-2011 21:56:30.565 [710]Demux : Audio to render 0.354 Sec
23-01-2011 21:56:32.988 [710]Demux : Audio to render 0.326 Sec
23-01-2011 21:56:36.737 [710]Demux : Video to render 0.750 Sec
23-01-2011 21:56:46.137 [710]Demux : Audio to render 0.263 Sec
Happens right after a switch to channel so i dont need to wait

 

[miroslav22]

I'm not 100% sure what those entries indicate (I don't think they're necessarily errors). I do see a lot of those recorded (via UNC or RTSP) without any problems with the picture or sound.

 

[romuz]

Yep but only till it do not come to 0.
But it is not important, main finding is step back fixes it and discontinuities

http://paste2.org/p/1207934

This shows the difference
its is viewing after switching and at 23-01-2011 22:05:45.080 i have done step back

 

[miroslav22]

Yep but only till it do not come to 0.
But it is not important, main finding is step back fixes it and discontinuities

Paste2: Next Generation Pastebin - Viewing Paste 1207934

This shows the difference
its is viewing after switching and at 23-01-2011 22:05:45.080 i have done step back

Ah right, so it seems to run ok once you've stepped back 15 seconds?

Apart from all the:

FileReader::Read() read to less bytes

 

[romuz]

Yes it runs perfect after skip

FileReader::Read() read to less bytes has no effect

 

[miroslav22]

Thanks that's a big help

Now it makes sense - it must be that tsreader is running too close to the live write point. The packets probably arn't available when it needs them (as they are appended much less frequently with the patch)

Hopefully i'll be able to change the code to replicate it now

 

[disaster123]

AW: Why oh Why Read() no open file

that would be perfect - any idea on how to solve this? Problem would be that channel switching is slower? Shouldn't Tswriter at least be one buffer ahead tsreader?

 

[miroslav22]

Hi guys,

Very sorry to disappoint but it doesn't look like i'm going to be able to fix the issues over UNC. I've tried everything I can think of but the initial stuttering and the file open problem I just cannot replicate.

I'm seeing all SD channels run quite well but some HD channels run very badly. There seems to be little difference in performance for me if the patch is used or not.

I've asked in the team forum for advice on the best way forward. The only solution I can propose is to have an option to switch tsreader/tswriter back to use the old method in the patch.

:sorry: