Inline.
Fabio Macchi wrote:
Greger, thanks for you answer, it helped to place order in my confused ideas.
Confusion at a higher level, that's what we all are striving for!!
About the bug you mentionated, has it been fixed in the cvs stable branch ( rel_0_8_14 , currently I have upgraded to this release ) or I have to install the main unstable branch ?
Yuo should at least download 0.9.x. ftp://siprouter.onsip.org/pub/gettingstarted/packages/ I think the fix is only in cvs head, but I'm not sure. Does anybody know? I think Jan did the fix.
I've sniffed a call from the SER interface connected to the GW: it appears that the entire SIP negotiation on this side ( I mean source IP address of UDP packet ) is between the GW ip ( x.x.x.243 ) and the SER IP assigned to the other interface (10.0.0.1 ) but, if I'm not wrong, I expected a SER UDP source address x.x.x.246 . Is this the problem ? May it depends from the mentionated bug ?
Well, if SER is running on two interfaces as well (and not only one public IP), then you need to turn on multihoming (mhome=yes) and you should probably turn on double record routing as well (see rr README) There's also another command in 0.9x where you can actually force a given interface. I have never used it.
force_send_socket()
After successfully SIP negotiation, I see RTP packet from GW -> rtpProxy ( x.x.x.244 -> x.x.x.246 ) and from client -> rtpproxy ( 10.0.0.2 -> 10.0.0.1 ), but rtpproxy is not bridging packets between the two interfaces.
Another question: is there any documentation about the use of force_rtp_proxy options more complete then nathelper README ?
That document only mentionate force_rtp_proxy("FI") and force_rtp_proxy("FE") ( hope this is correct in this case ), but in many example I see force_rtp_proxy("FII") or force_rtp_proxy("FIE") and I don't understand the double option.
Not that I know of. However, I searched through my archives and found a message with a working bridge scenario. For some reason, I don't think it was put at onsip.org. With the upcoming new iptel.org site, adding how-tos will be much easier for everybody :-) Meanwhile, I'll forward the message to the list. g-)