Testing plugins sticks on "Publishing..."

[Venelin Spiridonov]

Shortly - running Janus in docker on m1 Mac.

EchoTest for example sticks on "Publishing..." and exit with "{janus: "hangup" ...reason: "ICE failed"}".

Applied conf for STUN in janus.jcfg like: 

stun_server = "stun2.l.google.com"
stun_port = 19305
nice_debug = true

Janus boot log:

Initializing ICE stuff (Full mode, ICE-TCP candidates disabled, half-trickle, IPv6 support disabled)
STUN server to use: stun.l.google.com:19302
  >> ***.177.***.127:19302 (IPv4)
Testing STUN server: message is of 20 bytes
  >> Our public address is **.**.**.108

EchoTest still sticks on "Publishing...".

q1 - should I expect to see some reply from the STUN server if it is contacted OK?

q2 - am I missing something?

[Lorenzo Miniero]

ICE Failed means local and remote candidates couldn't connect to each other. You should inspect the related handle with the Admin API to check what's wrong:

https://www.meetecho.com/blog/understanding-the-janus-admin-api/

or check something like chrome://webrtc-internals or about:webrtc to check the candidates matrix. Good chances are you either only have private addresses on both ends that can't communicate, or something that makes it harder (weird NAT, firewall).

L.

[Venelin Spiridonov]

"local-candidates": [

                "1 1 udp 2015363327 10.1.0.21 44689 typ host",

                "2 1 udp 1679819007 77.78.***.*** 49496 typ srflx raddr 10.1.0.21 rport 44689"

            ],

            "remote-candidates": [

                "3573973069 1 udp 2122260223 192.168.0.102 51921 typ host generation 0 ufrag 20Dq network-id 1 network-cost 50",

                "3573973069 1 udp 2122260223 192.168.0.102 51644 typ host generation 0 ufrag 20Dq network-id 1 network-cost 50",

                "1414432505 1 udp 1686052607 77.78.***.*** 51921 typ srflx raddr 192.168.0.102 rport 51921 generation 0 ufrag 20Dq network-id 1 network-cost 50",

                "1414432505 1 udp 1686052607 77.78.***.*** 62303 typ srflx raddr 192.168.0.102 rport 62303 generation 0 ufrag 20Dq network-id 1 network-cost 50",

                "1414432505 1 udp 1686052607 77.78.***.*** 51644 typ srflx raddr 192.168.0.102 rport 51644 generation 0 ufrag 20Dq network-id 1 network-cost 50",

                "3573973069 1 udp 2122260223 192.168.0.102 62303 typ host generation 0 ufrag 20Dq network-id 1 network-cost 50"

            ],

This is what I've got after enabling AdminAPI.

TURN has been configured openrelay project - doesn't helps.

Above I see local and remote candidates with my external IP (77.78.***.***).

In the same moment I realised that I have Falcon on my Mac and probably this situation comes from it. Will look how to disable it and will see how it goes.

Anyway - good lesson learned with AdminAPI :).

[Venelin Spiridonov]

Decided to fight a bit more before blaming the firewall for all.

After heavier attack over the logs, I saw that remote tracks are coming (see attached) and after this the ICE fails.

If the local/remote parts can't see each other - is it possible to see such logs for incoming media?

[Lorenzo Miniero]

If ICE fails there's no media at all.

L.

[Venelin Spiridonov]

nat_1_1_mapping to public router address resolves the issue.