[OpenSER-Users] Intervention in provisional responses of stateful INVITE processing.

[Alex Balashov]

Daniel-Constantin Mierla wrote:

> Hello,
> 
> On 07/25/08 03:07, Alex Balashov wrote:
>> Greetings,
>>
>> An OpenSER REQUEST-ROUTE that statefully t_relay()s an INVITE request 
>> will automatically pass back a provisional "100 Trying" response to 
>> the sender.
>>
>> The problem I have is that I have a database operation that occurs in 
>> my main REQUEST-ROUTE that can potentially, under certain unfavourable 
>> circumstances, be somewhat latent.  This operation must run before the 
>> relaying can occur, since its results are used to rewrite the RURI, etc.
>> When the operation lags, sometimes it exhausts the T-timers and 
>> results in retransmission of the INVITE, or, for equipment configured 
>> in a particularly conservative way, even a failure.
>>
>> So, what I need is some way to pass a provisional "100 Trying" at the 
>> beginning of the REQUEST-ROUTE immediately in order to placate the 
>> sender.
>>
>> I can't just reply statelessly, because I need the stateful processing 
>> for other things related to call accounting, so I must use 'tm.'  And 
>> if I send a stateless "100 Trying" reply, the subsequent t_relay() 
>> will not know that and duplicate it, and although as far as I know 
>> this does not violate the RFC, it is still a situation I would like to 
>> avoid in principle.
>>
>> What I need to know is if there is some way in 'tm' to disable 
>> automatic provisional responses in a temporary or ephemeral fashion, 
>> or else to disable them entirely and then re-enable them later.
>>
>> Thank you so much in advance for all of your help!
>>   
> relay functions from tm take a parameter that can disable sending the 
> provisional reply.
> 
> http://www.openser.org/docs/modules/1.3.x/tm.html#AEN345

Hah.

"do not generate an 100 trying provisional reply when building the 
transaction. By default one is generated. Useful if you already pushed 
an stateless 100 reply from script."

Thank you, Daniel-Constantin.  From this, the only viable and necessary 
conclusion to draw is that I'm a flat-out idiot for completely 
overlooking this plainly obvious-as-day fact in what I imagined to be 
effective RTFMing.  :-)

-- 
Alex Balashov
Evariste Systems
Web    : http://www.evaristesys.com/
Tel    : (+1) (678) 954-0670
Direct : (+1) (678) 954-0671
Mobile : (+1) (706) 338-8599