[Serusers] Tags, forwarding, redirects, and picky UACs

Alistair Cunningham acunningham at integrics.com
Tue Apr 12 11:08:48 CEST 2005


Bogdan-Andrei,

Can you please point me to where in the RFC says this? I suspected this 
myself, did some reading, and came up with the following from section 
8.2.6.2:

"The same tag MUST be used for all responses to that request, both final 
and provisional (again excepting the 100 (Trying))."

Alistair Cunningham,
Integrics Ltd,
+44 (0)7870 699 479
http://integrics.com/


Bogdan-Andrei Iancu wrote:
> According to RFC 3261, the UAC *must* drop any To-tag received in a 
> provisional reply in favour of the To-tag advertised in the final reply 
> - be this it means not only to accept a final reply with a different 
> To-tag, but also to use it in generating sequential messages (like ACK, 
> re-INVITEs, etc)
> 
> So you should report the problem to your UAC vendor. Meanwhile you can 
> do a trick in SER cfg - use a reply_route to remove the To-tag from all 
> provisional replies....this will make happy the UAC.
> 
> bogdan
> 
> Alistair Cunningham wrote:
> 
>> I'm trying to implement forwarding to voicemail with a UAC that's picky
>> about the tags in the To: headers it receives.
>>
>> First, I tried the following:
>>
>> failure_route [1] {
>>     rewritehostport("192.168.1.1:5060");
>>     append_branch();
>>     t_relay();
>> }
>>
>> This failed because the UAC received a "180 Ringing" from the ringing
>> Grandstream phone via ser that contained a To: tag. After the timeout,
>> it received another 180 from Asterisk (the voicemail system) with a
>> different tag. Since it doesn't support forked INVITES, it dropped the 
>> call.
>>
>> Then I tried:
>>
>> failure_route [1] {
>>     rewritehostport("192.168.1.1:5060");
>>     t_reply( "302", "Moved Temporarily" );
>> }
>>
>> This time it receives the 180 from the phone as before. Then it receives
>> the 302. Alas, SER has ignored the tag from the phone, and created a new
>> one of it's own, so the UAC again drops the call.
>>
>> Does anyone know of a way round this? Can SER be persuaded to use a tag
>> it has already received? Can I strip the tag from the To: header in 
>> one or both of the responses? I'm willing to get my hands dirty with 
>> the SER source code if necessary.
>>
> 
> 
> 




More information about the sr-users mailing list