18 Dec
2014
18 Dec
'14
8:05 p.m.
On 12/18/14 12:55, Andrey Utkin wrote:
2014-12-18 19:30 GMT+02:00 Richard Fuchs
<rfuchs(a)sipwise.com>om>:
Amazon NAT is exactly why I've mentioned it, because on an Amazon
system, if you don't use the --interface option correctly
($INT_IP!$EXT_IP notation), you get exactly these kinds of write errors
that show in your log.
But these configs and this CLI line don't match the logs you posted
earlier. Is it also write errors you're getting with those?
cheers
Write error on RTP socket usually indicates an
incorrect network setup,
for example trying to use a source address for IP packets which isn't
bound to any local network interface (especially if you're sitting
behind NAT), or local iptables rules rejecting outgoing IP packets, or
missing IP routes to the destination. Perhaps post your network setup
and the CLI arguments to rtpengine you're using.
I use VPS from DigitalOcean. Likely no NAT. Only one public IP, and
one domain name used (whdd.org). Exactly the same situation was with
Amazon VPS behind amazonish NAT.
My configs are here: https://github.com/krieger-od/webrtc_kamailio_etc
(only kamailio.cfg and modules.cfg are used, kamailio*.cfg are not).
Also there you can check command line used to run rtpengine. To put it
there, it is
/home/krieger/rtpengine/daemon/rtpengine --interface=188.226.241.236
--listen-ng=127.0.0.1:22222 -m 30000 -M 35000 --foreground
--log-stderr