Hi all,
To prevent of a previous crash of my encoder or something like that, I would like, when I start my encoder to be sure that the publishing point is in proper state. No old streams currently started... No matter of previous data, i just want to restart from scratch.
So, i try to send a EOS signal (empty mfra box) as a first request before pushing my stream. My publishing point is setup with the "restartOnEncoderReconnect=true". When i simulate a crash, the publishing point stays in "Starded state" and when the encoder reconnects, it sends the emty mfra and the publihing point goes to "stopped state" and automatically restarts when new request arrives. So in this case, all is fine...
My problem is the following one : if i manually restart the publishing point when my encoder is not running. The PublishingPoint goes to the "Starting sate". When my encoder starts and sends the empty mfra, the publishing point is still in "Starting state", it never reachs the "Stopped state". I believe i arrives in a not managed state where my Publishing point is "starting" and my stream is stopped. (I saw in Incoming streams summary that there is only one total stream which is currently stopped). My encoder only uses a unique StreamdId. So i true believe that IIS refuses new fragments with this is streamId, because this stream is considered as stopped...
What can i do in this case ? Maybe i can force the "started state" by sending a "fake" stream to be sure to go to "Starting state" point before sending EOS, and after that, really pushing my stream... Is there any proper solution ?
Thanks in advance.
Anthony