it has been a while since we had this issue.
In a Room with 2 Publishers, Publisher A receives stream correctly from
Publisher B but B gets a black screen from A.
After browser refreshes from both parties, the scenario turned around.
Publisher B receives stream correctly from A but A gets black screen from B.
In a pool of many “Licode sessions” we monitored in the past 6 months, the
“black screen” issue didn’t appear until now.
The major difference between the successful Licode sessions and the failing
session mentioned above is distance (transatlantic East USA <-> West EU).
I’ve attached the logs of the failing session as attachments.
Both peers are using latest version of Chrome. A on Windows 8.1 and B on
We have a TURN server set up and relay candidates show up in both logs.
I have a gut feeling that the server we run the TURN server on might have
network performance issues when routing streams over TURN to Erizo.
We run it on a t1.micro on AWS EC2.
I will upgrade the EC2 instance to something better and try the session
In the meanwhile, it would be nice if the logs in attachment could be
doublechecked by someone with better insights to exclude any other problem
lurking in the setup.