Creating ICE agent and is attempted twice (and fails)

[sol....@rise.tech]

For some reason subscriptions in our test setup stopped working today. We are seeing this Error on the client side:

Ooops: 490 Error setting ICE locally
error creating answer 490 Error setting ICE locally

and the matching error in the Janus logs:

[Fri Jun  1 15:21:27 2018] [8206877793709622] Creating ICE agent (ICE Full mode, controlling)
[Fri Jun  1 15:21:27 2018] [WARN] [8206877793709622] Agent already exists?
[Fri Jun  1 15:21:27 2018] [ERR] [janus.c:janus_process_incoming_request:1179] Error setting ICE locally

So it seems that the first line there matches with a successfull call to janus_ice_setup_local, while the warning and error comes from a failed call (ice.c:2906).

As a result subscribers get no remote candidates and the stream just stops before it started; in chrome://webrtc-internals we see all the right local candidates and an 'icegatheringstatechange > complete' but no connection happening.

We have tried this with janus built from today's master branch, v0.4.1 and v0.3.1 and we are getting the same error in any case.

This happens both with google's STUN and our own coturn-backed STUN/TURN servers as iceServers.

Is there some known problems that could lead to the double-call of janus_ice_setup_local?

[Mirko Brankovic]

Which version of libnice do you have?

Do you see stun discovery packets being sent out?

I'm running 0.4.1 janus with libnice from deb repo over 2.4 libwebsockets, clients are mostly on same network but I didn't experienced this problems.

--
You received this message because you are subscribed to the Google Groups "meetecho-janus" group.
To unsubscribe from this group and stop receiving emails from it, send an email to meetecho-janu...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Lorenzo Miniero]

Are you sure you're not sending the SDP twice yourself? Or have you verified if the stack is still clearing itself when you try again? A few days ago we fixed an issue where the ICE loop was not being removed, and this caused the WebRTC stack never to be cleared, and so cause those issues. I'm not aware of such problems in master right now.

L.