Closed mattdsteele closed 3 years ago
You're right, that's a problematic situation at the moment. I do have it marked as a TODO
in the sources, as I'm not quite sure yet how I want to handle this situation:
Feedback appreciated!
I've opted for the second option: we try to re-establish a connection, but won't buffer any packages to keep the multiple outgoing streams in sync.
Fixed in master.
I think you're tracking on this, but I'm encountering broken pipes when there's a disconnect from the downstream RTMP streams (for example, if an Owncast instance restarts).
It manifests itself with a wall of errors like:
My current workaround is just to kill the prism instance.