You could turn on the debugging to see how ex. the avpairs are changed...
My first guess would be to re-verify that you really are using domains in usrloc
(parameter) and avp_pushto (ruri and not ruri/username). Besides that I cannot see
anything obvious. I have basically the same setup running fine.
If I interpret the code correctly, lookup() will use the received field as dst
(printed when doing a serctl ul show), if not, it will use ruri. The save command will set
the received field to either the contact ip:port or the avpair set by
modparam("registrar", "received_avp",def:42)
g-)
---- Original Message ----
From: Java Rockx
To: serusers
Sent: Monday, June 20, 2005 10:59 PM
Subject: [Serusers] avp_pushto() + append_branch() not working
Hi All.
I just need a sanity check regarding avp_pushto() for serial forking.
Someone please correct me if I'm wrong here.
When processing an INVITE message which results in a 486 Busy status
my failure_route[] gets invoked.
In the failure route I use avp_pushto() to load the next leg of the
call in to the $ruri and then call append_branch() and
lookup(location) prior to t_relay()ing the message.
Using ngrep I do see the R-URI being changed to the value used in the
avp_pushto() call, but the IP address that the INVITE is then sent to
is the same as the IP address of the SIP router, which I cannot
understand because I do call lookup("location") before t_relay()
In otherwords, if my sip router,
sip.mycompany.com has an IP of
71.72.73.74 and I use avp_pushto() to set the R-URI to
1001_at_sip.mycompany.com, lookup(location) does return TRUE, but
ngrep still shows the INVITE leaving the SIP router and going to
71.72.73.74 (ie, itself) rather than the IP address in usrloc.
Is there anything obvious that I'm not doing that would cause
lookup(location) to return TRUE, but not change the R-URI
accordingly?
Regards,
Paul
_______________________________________________
Serusers mailing list
serusers(a)lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers