did something change in how the audiobridge does RTP forwards?

105 views
Skip to first unread message

august black

unread,
Jun 24, 2022, 5:04:03 AM6/24/22
to meetecho-janus
Hi,

I've had great success in the past with starting an rtp forwards in the audiobridge and then creating an sdp for ffmpeg to read.

Given an rtp forward on port 10040, I'd make an sdp like so:

v=0
o=- 0 0 IN IP4 127.0.0.1
s=FFmpeg
c=IN IP4 127.0.0.1
t=0 0
m=audio 10040 RTP/AVP 111
a=rtpmap:111 opus/48000/2
a=sendonly

and then read it with ffmpeg like so

ffmpeg -protocol_whitelist file,udp,rtp -f sdp -i my.sdp ....


I now get a timeout error from ffmpeg like it cannot read the RTP stream from janus.

It was working great for me up until recently where I changed two things.  I upgraded janus and I reconfigured how I run both  janus and ffmpeg.

Would the upgrade of janus from x.11 to 0.x be meaningful?  I wouldn't think so.

How about the docker setup?

Previously, I'd run both the janus and the ffmpeg process each on separate docker images using "host" networking in a docker-compose setting.   That worked fine.  However, now I am running only janus in a docker container with "host" networking and the ffmpeg directly on the host.  That is where I see the timeouts as if ffmpeg cannot read the RTP stream from janus.

Is there maybe a janus config setting that stops RTP quietly?  Is there maybe some host setting that needs to be set for janus to open the RTP port properly?

thanks for any advice -August.





Lorenzo Miniero

unread,
Jun 24, 2022, 6:12:11 AM6/24/22
to meetecho-janus
You can check the changelog to see the changes since the version you were using:

I see a couple of mentions of AudioBridge forwarders in there, in case any of those changes is to blame for the issue you have. You can also try using git bisect between the latest 0.x and a commit that worked for you to nail down the commit that oyu think broke it.

L.

august black

unread,
Jun 26, 2022, 3:41:05 PM6/26/22
to Lorenzo Miniero, meetecho-janus

Something very subtle seems to have changed.

If I create the rtp forward on 'localhost', ffmpeg doesn't get any input.

However, if I create the rtp forward on '127.0.0.1', it all works fine.

Should that be the case?

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/meetecho-janus/1f02d2d6-d39e-4b46-adb2-0aa22e0546can%40googlegroups.com.

Lorenzo Miniero

unread,
Jun 27, 2022, 4:21:51 AM6/27/22
to meetecho-janus
That's always been the case, and it's related to how localhost is resolved. Sometimes it resolves to "::1" and so if you're bound to 127.0.0.1 you won't get it. Happens to me as well, always has. Just bind to 127.0.0.1 if that's what you want, or force the family to ipv4 when creating the forwarder.

L.

august black

unread,
Jun 27, 2022, 5:38:14 AM6/27/22
to Lorenzo Miniero, meetecho-janus

Okay, I think I see now. 

How localhost is mapped will depend on the environment.

I think that happened because I installed IPv6 on my digital ocean droplets where I run Janus

Thanks for the support.



Reply all
Reply to author
Forward
0 new messages