Ah, I see.. I totally forgot about an internally generated 408. I was just trying to suggest that if you receive a 408 from an endpoint, it isn't necessarily grounds for being blacklisted. However, given the application, it probably makes sense.
On Wed, Feb 11, 2009 at 5:04 PM, Iñaki Baz Castillo <[email protected]> wrote: > El Miércoles, 11 de Febrero de 2009, Brett Nemeroff escribió: > > That 408 doesn't make sense to me.. > > > > If you got a 408 back, the gateway you are sending to WORKS, but IT'S > > destination is failing. > > That's incorrect. RFC 3261 states clearly that, in case a request has no > reply > (after the corresponding retransmissions) then the transaction client > (OpenSIPS in this case) MUST generate *internally* a 408 response. In case > there is just one branch, OpenSIPS must send that 408 upstream as > definitive > negative response. And sure this is the behaviour. > > > > say for example you send to a gateway and that gateway sends to 10 > > different providers.. if this call routes to provider A, which doesn't > > respond, then your gateway may respond back with a 408. Where the failure > > isn't your gateway, it's your gateway's upstream.. > > That's a more complex scenario. But you are rigth, it's not easy to choose > the > appropiate response code since any node in the signalling path could > generate > that response. > > > > -- > Iñaki Baz Castillo > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users >
_______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
