Closed
Bug 1250202
Opened 9 years ago
Closed 7 years ago
One-way call with talky.io when one end was in a Hotel network
Categories
(Core :: WebRTC: Networking, defect, P1)
Core
WebRTC: Networking
Tracking
()
RESOLVED
INCOMPLETE
backlog | webrtc/webaudio+ |
People
(Reporter: mreavy, Assigned: bwc)
Details
Attachments
(3 files)
Adam and I tried to have a call while he was in a hotel. We had one-way media - I could see and hear him perfectly; he could not hear or see me.
Reporter | ||
Comment 1•9 years ago
|
||
P1 15 to at least analyze what we captured
backlog: --- → webrtc/webaudio+
Rank: 15
Priority: -- → P1
Reporter | ||
Comment 2•9 years ago
|
||
Reporter | ||
Comment 3•9 years ago
|
||
Reporter | ||
Comment 4•9 years ago
|
||
Assignee | ||
Updated•9 years ago
|
Assignee: nobody → docfaraday
Assignee | ||
Comment 5•9 years ago
|
||
I'm not totally sure what is going on here (the pcap is from a call that wasn't in mreavy's about:webrtc), but it seems to be related to the handling of remote peer reflexives when using a relay. In experimenting with the test TURN server, I've found that Firefox is doing a very poor job of setting up TURN permissions; it is possible that no permission was set up for abr's peer reflexive, resulting in abr's packets being discarded.
Reporter | ||
Comment 6•9 years ago
|
||
Byron -- I thought that the about:webrtc I uploaded had everything (the info from the earlier calls as well as the info from the last one where abr took the pcap). Are you sure it's not there? If you can do a quick sanity check, I'd appreciate it.
Assignee | ||
Comment 7•9 years ago
|
||
I think there is an overlap between the two about:webrtc, but the pcap contains only packets from one call, which is not in your about:webrtc.
Assignee | ||
Comment 8•9 years ago
|
||
Looking at the packet capture, I'm definitely seeing media arriving.
Assignee | ||
Comment 9•9 years ago
|
||
abr's about:webrtc seems to indicate that media was arriving too. Was this a normal call that went one-way after a while?
Flags: needinfo?(mreavy)
Comment 10•9 years ago
|
||
(In reply to Byron Campen [:bwc] from comment #9)
> abr's about:webrtc seems to indicate that media was arriving too. Was this a
> normal call that went one-way after a while?
No, I never saw or heard Maire. She was able to see me from the very beginning of the call.
We did run several calls, and the situation was 100% reproducible. It's possible that the attached pcap is from a later call than the about:webrtc traces. The symptoms were always identical, however.
If you need to gather additional information, the hotel is actually located close to some of our friends, and I've confirmed their willingness to run through diagnostics as necessary (likely on Fridays only, though). Let me know if you'd like to put in touch.
Updated•9 years ago
|
Attachment #8722036 -
Attachment mime type: text/plain → text/html
Assignee | ||
Comment 11•9 years ago
|
||
Hmm, on closer inspection, I have a pcap from abr's side, but no about:webrtc for that call, and about:webrtc from maire's side. The pcap has bidirectional media, it appears.
Reporter | ||
Comment 12•9 years ago
|
||
Looking at abr's about:webrtc log (which was probably from the earlier call), there was incoming media data to Abr. I think getting another capture with more logging turned on would be very interesting. Tomorrow is Friday; so if we want another capture, we should ask them today.
Flags: needinfo?(mreavy)
Updated•7 years ago
|
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → INCOMPLETE
You need to log in
before you can comment on or make changes to this bug.
Description
•