[OpenSIPS-Users] DB-ONLY mode call between two proxies failed.

[osiris123d]

My question was referencing this email thread


Hi Jiang, 

the problem is the USRLOC module saves (in the user location record) the 
socket the REGISTER was received from (to use it when sending the 
traffic to the users). But if you have multiple servers on the same DB, 
there will be a kind of conflict between the stored sockets (record was 
inserted by R1 with socket S1 and later R2 reads the records, R2 will 
try to use S1 which obviously is not local). 

I can send you a small patch do get rid of that, but if you are dealing 
with NATed clients, it will not works - a natted client can be handled 
only by the same server. 

Regards, 
Bogdan 

Jiang Jinke wrote: 

> I'm using the same location table with two/three proxy server, but I 
> can't make calls between the proxies. 
> Do I have to do replication in SIP level ? I'd like to do this in a 
> simple way, but if replication will make the arch more flexible, then 
> it's acceptable with a more complicate config. 
> 
> I'm having the following message in my syslog when the other endpoint 
> is not in the same proxy server which was receiving the call. 
> 
> syslog content: 
> WARNING:usrloc:dbrow2info: non-local socket
> <udp:xx.xx.xx.xx:5060>...ignoring 
> ERROR:core:get_out_socket: no socket found 
> ERROR:tm:update_uac_dst: failed to fwd to af 2, proto 1  (no 
> corresponding listening socket) 
> ERROR:tm:t_forward_nonack: failure to add branches 
> 
> Regards, 
> Jinke Jiang 

-- 
View this message in context: http://n2.nabble.com/DB-ONLY-mode-call-between-two-proxies-failed-tp3926843p4801720.html
Sent from the OpenSIPS - Users mailing list archive at Nabble.com.