[OpenSIPS-Users] RTPProxy bridge mode issue

[Yuri Kirsanov]

Hello,

I have an RTP Proxy running in bridge mode.
I've got a call coming from LAN side of RTP Proxy and outgoing to WAN
(Internet) side of RTP Proxy. I'm using rtpproxy_offer("fioc") command if
nat_uac_test("4") believes that call comes from LAN and
rtpproxy_offer("feoc") if it comes from WAN. RTPProxy is run in bridge mode,
I've specified both internal and external IP addresses. But here's what's
going on on a call coming from LAN:

Aug 17 13:52:44 localhost rtpproxy[17736]: DBUG:handle_command: received
command "20370_72 UI
6d8ab1d978f708b918f5337b24a3359a at 10.65.0.3:506010.65.0.3 16972
as4bcd2ee5;1"

Aug 17 13:52:44 localhost rtpproxy[17736]: INFO:handle_command: new session
6d8ab1d978f708b918f5337b24a3359a at 10.65.0.3:5060, tag as4bcd2ee5;1 request
ed, type
strong

Aug 17 13:52:44 localhost rtpproxy[17736]: INFO:handle_command: new session
on a port 63626 created, tag as4bcd2ee5;1
Aug 17 13:52:44 localhost rtpproxy[17736]: INFO:handle_command: pre-filling
caller's address with 10.65.0.3:16972
Aug 17 13:52:44 localhost rtpproxy[17736]: DBUG:doreply: sending reply
"20370_72 63626 10.65.0.254#012"

So, all good here, RTP Proxy got "UI" command meaning "Create new session
using Internal IP" and used correct one - 10.65.0.254.

And then call's being answered and I'm again doing same nat_uac_test("4") to
check if it's local or not, it determines correctly that call is coming from
the Internet, issues command rtpproxy_answer("feoc"), but in logs I see
following:

Aug 17 13:52:48 localhost rtpproxy[17736]: DBUG:handle_command: received
command "20375_4 LE 6d8ab1d978f708b918f5337b24a3359a at 10.65.0.3:5060 202.129.
XXX.XXX 7078 as4bcd2ee5;1
671437651;1"

Aug 17 13:52:48 localhost rtpproxy[17736]: INFO:handle_command: lookup on
ports 63626/59248, session timer restarted
Aug 17 13:52:48 localhost rtpproxy[17736]: INFO:handle_command: pre-filling
callee's address with 202.129.XXX.XXX:7078
Aug 17 13:52:48 localhost rtpproxy[17736]: DBUG:doreply: sending reply
"20375_4 59248 10.65.0.254#012"

So, "LE" command received by proxy from OpenSIPS meaning that External IP
address has to be used, client's external IP address specified, BUT RTP
Proxy for some reason replies with INTERNAL IP and expects packets on that
interface!!!

Do you have any idea why is it behaving this way? I'm using OpenSIPS 1.6.4
and RTP Proxy 1.2.1 without any patches. I'm not using "mhomed=1" option,
I'm using "force_send_socket()" on outgoing calls from Internet client to
LAN and it works fine. Should I also use that command for outgoing calls
from LAN to Internet client? But that shouldn't affect RTP Proxy behavior,
shouldn't it? Also, I've tried to use "r" and "w" options when invoking RTP
Proxy, but that doesn't help.

Thanks and best regards,
Yury Kirsanov.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20110817/5e514a35/attachment.htm>