[OpenSIPS-Users] INVITE not forwarded, call fails

[Bogdan-Andrei Iancu]

Or maybe the INVITE from the biascica to yhe asymmetric is not going 
through because of the NAT (maybe the client is doing a poor job in 
keeping the pinhole opened in the nat). So, as time as the call is 
originated from behind the nat, it will work, but when  from outside, it 
will not be able to traverse the nat and reach the client.

Regards,
Bogdan

Raúl Alexis Betancor Santana wrote:
> On Sunday 08 November 2009 15:47:38 lorenzo wrote:
>   
>> On 04/11/09 13:41, lorenzo wrote:
>>     
>>> On 04/11/09 13:32, Iñaki Baz Castillo wrote:
>>>       
>>>> El Miércoles, 4 de Noviembre de 2009, lorenzo escribió:
>>>>         
>>>>> and this is the wireshark capture of the "conversation":
>>>>> http://pastie.org/683069
>>>>>           
>>>> this shows a call which is cancelled (CANCEL) and later a call which
>>>> receives 180 Ringing,
>>>>
>>>> It's difficukt to inspect  trace if you don't say what exactly occurs in
>>>> that trace,
>>>>         
>>> hi Iñaki, thanks for your interest!
>>>       
>> guys, nobody got any advice on where to look for to solve this problem?
>> if anybody needs more info, just let me know!
>>     
>
> What I see from you trace, (better if next time you put a ngrep-sip trace ..), 
> is:
>
> Apple                         DLink                          GW
> INVITE      ->
>                  <-         100 Trying
> CANCEL    ->
>                  <-         200 Canceling
>                  <-         486 Request Canceled
> ACK           ->
> -------------
> New INVITE ->
>                     <-      100 Trying
> [the next is supposed, not in trace]
>                               INVITE              ->
>                                                       <-      100 Trying
> [... end supposing]
>                                                        <-     INVITE
>                     <-      INVITE
> 180 Ringing ->
>
>
> Maybe you are rerouting the second invite back to your UAC, or you do some 
> aliasing ... or something similar.
>
>