Transcode stop but rmtp publish is still running

Hi there!

I absolutely love this project. Was a life saver for my project.

First, here is my use case: I have an electron app that uses FFMPEG to send IP camera streams to Node-Media-Server hosted on a EC2 instance. I am using transcoding to generate HLS files that users can play on a mobile app and website.

On the electron app I have a routine that restarts ffmpeg command if it exits. On Node-Media-Server I have used the events to notify the viewers when a stream is available or offline.

Now the issue: sometimes Transcode just stop when the publishing is still happening. That way my app can’t restart FFMPEG because it is still running even though the transcoding is not.

When the transcoding stop this was the log displayed:

3/16/2021 01:55:43 1 [INFO] [rtmp play] Close stream. id=2ZRT4FJT streamPath=/live/6048f9e6d650ab0007d1ee6d streamId=1

3/16/2021 01:55:43 1 [INFO] [rtmp disconnect] id=2ZRT4FJT

3/16/2021 01:55:43 1 [INFO] [Transmuxing end] /live/6048f9e6d650ab0007d1ee6d

Any help or guidance will be much apreciated.

[flavioribeirojr]

I have decided to migrate to https://github.com/arut/nginx-rtmp-module

So far I have no issues.

[benjamin658]

For anyone who is struggling with this issue, we finally found that in the document config, ping and ping_timeout should be exchanged:

const config = {
  rtmp: {
    port: 1935,
    chunk_size: 60000,
    gop_cache: true,
    ping: 60, // 30 -> 60
    ping_timeout: 30 // 60 -> 30
  },
  http: {
    port: 8000,
    allow_origin: '*'
  }
};