[Ekiga-list] STUN registration

Wolfram Wagner Wolfram.Wagner at web.de
Sat Dec 23 20:01:47 UTC 2006

Hello Damien, 

Thanks for your reply. It took a while until I had time again.

Please remember, I have problems with registration and get a 404 not 
found message.
 Last time you answered me, that I should sniff a bit. So I did.

I found out, that registration at STUN server is done maybe by using a 
wrong IP. My current IP is

In STUN Registration I see contact binding:
	URI: <sip:wolfram.wagner at;transport=udp>

Then my client sends SUBSCRIBE using the same wrong IP Address.

Next I get a 404 Error (not found) telling me 
Sent-by Address:
Sent-by port: 5074
RPort: 5074
Warning: 392 "Noisy feedback tells:  pid=16986 
req_src_ip= req_src_port=5074 
in_uri=sip:wolfram.wagner at ekiga.net 
out_uri=sip:wolfram.wagner at ekiga.net via_cnt==1"

So I wonder where this IP comes from. I checked it and it 
is one of the dynamic addresses used by my provider. The check via 
ekiga.net/ip returnes the correct ip.

After three more Register retries it gives up.

Might it be possible that this wrong ip address makes problems in 
registration. Maybe the answer packages are sent to this machine.

Thanks and Have a very nice Christmas!


Your last answer:
Le vendredi 01 d?cembre 2006 ? 20:57 +0100, Wolfram Wagner a ?crit :

> I have posted it to  http://pastebin.ca/263742
> I see a 404 Error followed by a timeout when waitung for PDU. I have 
> enough insight to know what this means... 

The 404 is for the SUBSCRIBE (it is a normal behavior). The timeout is
for the REGISTER.

That means that :
- the SUBSCRIBE reaches ekiga.net
- the REGISTER, which is identical, doesn't reach ekiga.net OR the reply
to the REGISTER doesn't reach your internal host

Can you sniff the traffic on the routers to see what happens?

I have the feeling that one of them is either dropping the REGISTER, or
dropping the answer to the REGISTER or simply rewriting the REGISTER in
such a way that the answer never reaches you...

More information about the ekiga-list mailing list