What I see more as a possible obstacle here, which I also already mentioned is the usage
of two interfaces. In a TMS/VCS-C+E/Jabber deployment you have a clear separation
in between public and internal configuration parameters as well as the handling on the client.
On the starter pack I at least do not recall having seen any representation of that, I think
it always tries to register to the clustername.
Thats why I recomend to use a one interface setup in a DMZ, so the internal clients
can reach that IP as well as the external ones.
Regards the log, it would be interesting to have the log containing the INVITE and 200 OK to the invite
to see the ips used for media, best would be for an internal (working) and an external (non working) call.
And this for both call legs received from and send to the clients.
The Log level should be Debug not trace which makes it easier to read and still contains enough info.
Such logs are also better attach a plain text or some online clipboard service file rather than
posting it in the message. It gets nearly unreadable here with so long texts, ...
Anyhow if Hugo does not see progress or we run out of ideas here or there is no direct answer from Cisco
he could escalate this thread as a service request.
