You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I originally discovered the problem using LibDataChannel with libjuice but it can be reproduced with just libjuice.
It seems like libjuice wrongly filters out a packet coming from the TURN server because the port number does not match, because of the symmetric NAT.
I've tried to fix the filtering issue, but then the rendezvous still fails as the ip:port tupple is not added to the list of candidates and the rendezvous still fails.
The text was updated successfully, but these errors were encountered:
Thank you for reporting. Could you please clarify two things:
What port number does not match exactly? NATs translate the source address of outgoing packets and the destination address of corresponding incoming packets, so packets sent from the TURN server to an agent (i.e. incoming packets) can't have the source address changed.
libjuice does not filter by port number. What have you fixed exactly?
I have discovered an issue when running the TURN test while the machine performing the test is behind symmetric NAT (wifi teethering from my phone) https://github.com/paullouisageneau/libjuice/blob/master/test/turn.c
I originally discovered the problem using LibDataChannel with libjuice but it can be reproduced with just libjuice.
It seems like libjuice wrongly filters out a packet coming from the TURN server because the port number does not match, because of the symmetric NAT.
I've tried to fix the filtering issue, but then the rendezvous still fails as the ip:port tupple is not added to the list of candidates and the rendezvous still fails.
The text was updated successfully, but these errors were encountered: