Hello!
I'm having a issue with NAT and rtpproxy. Usually my setup works fine with natted clients, the Connection Information is overwritten with the IP of the rtpproxy and audio passes through in both directions. However, today i came across a problem where the Polycom 501 sets a outgoing ip of 0.0.0.0 instead of the private ip after resuming a call that was on hold(actually, the other party is invited again) - and the force_rtp_proxy ()-command on openser left the ip untouched instead of overwriting it with the rtpproxy-ip. As a result the person that was on hold had audio but the polycom user(with the "wrong" ip) hadn't.
The false ip left aside, is it expected behaviour of force_rtp_proxy to not touch 0.0.0.0?
Just out of curiosity - does someone know the "on hold"-problem with polycoms?
thx christian
Since the on-hold is signaled with 0.0.0.0, it must not be changed by nathelper. Maybe the polycom didn't toggle the on-hold. Make sure that you handle properly the re-INVITEs, since they follow loose_route(). Just check the following config, it may help abit:
http://voip-info.org/wiki/view/OpenSER+And+RTPProxy
Cheers, Daniel
On 10/13/06 20:47, Benko wrote:
Hello!
I'm having a issue with NAT and rtpproxy. Usually my setup works fine with natted clients, the Connection Information is overwritten with the IP of the rtpproxy and audio passes through in both directions. However, today i came across a problem where the Polycom 501 sets a outgoing ip of 0.0.0.0 instead of the private ip after resuming a call that was on hold(actually, the other party is invited again) - and the force_rtp_proxy ()-command on openser left the ip untouched instead of overwriting it with the rtpproxy-ip. As a result the person that was on hold had audio but the polycom user(with the "wrong" ip) hadn't.
The false ip left aside, is it expected behaviour of force_rtp_proxy to not touch 0.0.0.0?
Just out of curiosity - does someone know the "on hold"-problem with polycoms?
thx christian
Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users
Hello!
hmm, i believe i'm handling the re-INVITEs correctly, i've attached the relevant part of the config.
thx christian
On Sat, 14 Oct 2006 21:09:38 +0300 Daniel-Constantin Mierla daniel@voice-system.ro wrote:
Since the on-hold is signaled with 0.0.0.0, it must not be changed by nathelper. Maybe the polycom didn't toggle the on-hold. Make sure that you handle properly the re-INVITEs, since they follow loose_route (). Just check the following config, it may help abit:
http://voip-info.org/wiki/view/OpenSER+And+RTPProxy
Cheers, Daniel
On 10/13/06 20:47, Benko wrote:
Hello!
I'm having a issue with NAT and rtpproxy. Usually my setup works fine with natted clients, the Connection Information is overwritten with the IP of the rtpproxy and audio passes through in both directions. However, today i came across a problem where the Polycom 501 sets a outgoing ip of 0.0.0.0 instead of the private ip after resuming a call that was on hold(actually, the other party is invited again) - and the force_rtp_proxy ()-command on openser left the ip untouched instead of overwriting it with the rtpproxy-ip. As a result the person that was on hold had audio but the polycom user (with the "wrong" ip) hadn't.
The false ip left aside, is it expected behaviour of force_rtp_proxy to not touch 0.0.0.0?
Just out of curiosity - does someone know the "on hold"-problem with polycoms?
thx christian
Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users
There's an option for the Polycom phones to switch the hold behaviour.
Set voIpProt.SIP.useRFC2543hold to 0, and it should use RFC3264 rules for signalling hold instead of 0.0.0.0.
Benko wrote:
Hello!
I'm having a issue with NAT and rtpproxy. Usually my setup works fine with natted clients, the Connection Information is overwritten with the IP of the rtpproxy and audio passes through in both directions. However, today i came across a problem where the Polycom 501 sets a outgoing ip of 0.0.0.0 instead of the private ip after resuming a call that was on hold(actually, the other party is invited again) - and the force_rtp_proxy ()-command on openser left the ip untouched instead of overwriting it with the rtpproxy-ip. As a result the person that was on hold had audio but the polycom user(with the "wrong" ip) hadn't.
The false ip left aside, is it expected behaviour of force_rtp_proxy to not touch 0.0.0.0?
Just out of curiosity - does someone know the "on hold"-problem with polycoms?
thx christian
Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users
Hmm, actually this parameter is set to 0(default). However, in practice it still uses the old standard(rfc2543). There's no media direction parameter in the sdp-message sent by the polycom for some reason, although the manual states to do so (firmware 2.0.1). Is it possible to force this somehow?
regards christian
On Sun, 15 Oct 2006 19:58:33 -0400 Nathan Hawkins utsl@quic.net wrote:
There's an option for the Polycom phones to switch the hold behaviour.
Set voIpProt.SIP.useRFC2543hold to 0, and it should use RFC3264 rules for signalling hold instead of 0.0.0.0.
Benko wrote:
Hello!
I'm having a issue with NAT and rtpproxy. Usually my setup works fine with natted clients, the Connection Information is overwritten with the IP of the rtpproxy and audio passes through in both directions. However, today i came across a problem where the Polycom 501 sets a outgoing ip of 0.0.0.0 instead of the private ip after resuming a call that was on hold(actually, the other party is invited again) - and the force_rtp_proxy ()-command on openser left the ip untouched instead of overwriting it with the rtpproxy-ip. As a result the person that was on hold had audio but the polycom user (with the "wrong" ip) hadn't.
The false ip left aside, is it expected behaviour of force_rtp_proxy to not touch 0.0.0.0?
Just out of curiosity - does someone know the "on hold"-problem with polycoms?
thx christian
Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users
-
Benko -
Daniel-Constantin Mierla -
Nathan Hawkins