Transcoding service

[sub]

It looks like the request only arrived at the server at 19:23:37, and it responded immediately:

2016-08-24 19:23:37.919 [DEBUG][21] Got Web Request (172.16.3.62): /services/services method=channel.transcode.lease&sid=c25d129c7c614e66950847f746111b99
2016-08-24 19:23:37.919 [DEBUG][21] method=channel.transcode.lease
2016-08-24 19:23:37.919 [DEBUG][21] parameters:
2016-08-24 19:23:37.919 [DEBUG][21] method: channel.transcode.lease
2016-08-24 19:23:37.919 [DEBUG][21] sid: c25d129c7c614e66950847f746111b99
2016-08-24 19:23:37.919 [DEBUG][21] client_ip: 172.16.3.62
2016-08-24 19:23:37.919 [DEBUG][21] host_callback: 172.16.3.2:8866
2016-08-24 19:23:37.919 [DEBUG][21] physical_path: C:\Users\Public\NPVR\web\
2016-08-24 19:23:37.920 [DEBUG][21] user_agent: Python-urllib/2.7
2016-08-24 19:23:37.920 [DEBUG][21] range:
2016-08-24 19:23:37.920 [DEBUG][21] requested_path: /services/services
2016-08-24 19:23:37.920 [DEBUG][21] verb: GET
2016-08-24 19:23:37.920 [DEBUG][21] <?xml version="1.0" encoding="utf-8" ?>
<rsp stat="ok">
</rsp>

[mvallevand]

It looks like the request only arrived at the server at 19:23:37, and it responded immediately:

I just can't figure out where it sits for those six seconds and because of the5 seconds lease there is nothing I can do about it.

Also technically that call should fail since there is nothing to lease. I guess I have to a status check too after a long call.

Martin

[sub]

Actually, it takes a 10 seconds until it decides the device is no longer leased, to allow for clients that took a little too long.

[sub]

If you try using wireshark, can you see the request coming in earlier?

[mvallevand]

Yes I guess I could lease more frequently to give me more time

Martin

[sub]

How is the CPU usage at the time, if it's bogged down, maybe the webserver didn't getting cycles to full receive the request and start processing it?

[mvallevand]

How is the CPU usage at the time, if it's bogged down, maybe the webserver didn't getting cycles to full receive the request and start processing it?

High the CPU is always 75-98% but then again VNC is adding some CPU and network overhead while I am monitoring maybe I will set up perfmon

Edit: timing out and aborting on the lease call with enough time to reissue another lease seems to resolve the issue. One things I can do to trigger a disruption is reading the m3u8 file as it gets to be thousands of lines long. I you could avoid dumping it to the log that might help

Martin

[mvallevand]

Sub if you have a chance could you install this x-newa in your Kodi (16 is fine) http://forums.nextpvr.com/showthread.php...-0-Alpha-1, configure it to enable transcoding and let it run for 15 minutes on live tv and send the Kodi.log I'd like to any error messages with a fast backend.

Thanks

Martin

[sub]

Do you need debug logging turned enabled, or normal kodi.log is fine? I kicked off a streaming transcoded live tv a few minutes ago, and just realised I didn't have debug logging enabled in kodi.

I did notice with transcoded playback enabled, it starts with a still frame, and I had to hit the play button to force it to start playing.

[mvallevand]

Do you need debug logging turned enabled, or normal kodi.log is fine? I kicked off a streaming transcoded live tv a few minutes ago, and just realised I didn't have debug logging enabled in kodi.

I did notice with transcoded playback enabled, it starts with a still frame, and I had to hit the play button to force it to start playing.

No problem since I use python I control the prints to the log. I use Krypton and don't have a force play problem at the beiginning (it oftens plays twice) but I occasionally do hit a still frame and have to hit play m3u8 buffering isn't like ts with autoo-play. Currently I don't wait to 100% buffering but I haven't seen it under 100 which might be a problem at the beginning.

Martin