Good afternoon, plenty of your knowledge, since for more we look at the logs, and network connectivity, we do not see anything strange.
I commented, we create a Server or Origin application, we activate Httporigin,
We define
CupertinochalablesEgments True.
Cupertinocaccontrolplaylist at 10.
Cupertinocaccacontrolmediachunk 3600.
CUPERTININOMLEYLISTCHUNKCOUNT 6.
CupertinoConkduraTionGet 1000.
Cupertinomaxchunkcount 6.
CUPERTINOPLAYLISTCHUNKCOUNT 12.
We activate Transcoder to perform ABS (Apr) - H264 Main Coder Beamr
We generated switch.Smil with all three [Qualities, Source, 720p H264 Main 2MBITS and 360P H264 Main 1Mbits]
We emit a Live with VMIX to 3.1 Mbit / s
And we see how the WSM receives the signal and begins to transcode in the three qualities.
We open the RTMP in VLC to monitor that the signal is correct (the 3 qualities) is fluent and without stops.
.
We open the HLS URL by calling the smil generated with the three streams. (https://xxxxxstreamlok.net:1936/appxxx/smil:switch.smil/playlist.m3u8) In Different Player videoJS, HLS.JS, WOWZAPLAYER, Theoplayer On All Reproduces The URL, but we continuously receive cuts from the Chunk suffering from the emission. And we have to recharge the player.
Entire time Tests We carry out only with one or two clients simultaneously. and the result is continuous parons and 404 errors of chunkxxx.ts loss
The hardware used is VM 16 cores CPU, 60GB RAM, NFS storage connected by 4GBS.
The VM never passes from 30% of CPU, nor 40% RAM. Java is active in production mode 20000mb G1 (Garbage)
They could guide us where the cause can be. Since if we have the problems I am afraid that if I send this to a CDN, little podremso do.
Thanks a lot ,
All the best.