Hi Daniel,
Seems like I missed your reply.
I've tried with tcp only, and set_forward_no_connect is used also now.
The only log shown with debug=3 is, for the failing branch:
/usr/sbin/kamailio[60054]: [INVITE] ERROR: tm
[../../forward.h:240]: msg_send_buffer():
tcp_send failed
/usr/sbin/kamailio[60054]: [INVITE] DEBUG: tm [t_fwd.c:1524]:
t_send_branch(): t_send_b
ranch: send to 67.188.119.56:50239 (3) failed
/usr/sbin/kamailio[60054]: [INVITE] WARNING: tm [t_fwd.c:1543]:
t_send_branch(): ERROR:
t_send_branch: sending request on branch 0 failed
And after ( or before ), the logs shows sending to the other UA
registered with the same account, no mention of branch failure
route and t_relay replies true.
It seems like as there is one branch that was still connected (
tcp wise ), t_relay does not return a negative status, and the
branch failure route is never called.
That works as expected if both Contacts have lost their tcp
connection though ( t_relay returns a negative value ).
Hello,
have you tested only with tcp and set_forward_no_connect or also with udp?
If you run with debug=3, do you get any other hints about what happens there? Also, you can use debugger module with cfgtrace enabled to see what actions are executed.
Cheers,
Daniel
On 20/05/15 00:31, Tristan Mahé wrote:
Hi fellow kamailio users, I'm playing actually with t_on_branch_failure, in a simultaneous forking scenario ( lookup("location") returns multiple contacts ). Any idea on how to get those "on_branch_failure" working ? Thanks in Advance ! * Test by putting failure_exec_mode: - at 1 => no branch failure being called, never. - at 0 => t_relay does not fail if a single branch answers the transaction, even if all other branches are failing. * Test by moving the t_on_branch_failure("manage_branch_failure"); line before or after the lookup("location") or on a _on_branch route to set it per branch. - The on_route branch are executed, but never the failure one. - tried also with failure_exec_mode = 0 or 1. * Using set_forward_no_connect, I can see in the logs a "Sending to branch X failed" but no failure route being executed. Without it, nothing in the logs indicating that the branch has timeout. * Config extract: # auto-discard branches from previous serial forking leg modparam("tm", "failure_reply_mode", 3) # 3 by default # default retransmission timeout: 30sec modparam("tm", "fr_timer", 30000) # default invite retransmission timeout after 1xx: 120sec modparam("tm", "fr_inv_timer", 120000) modparam("tm", "reparse_on_dns_failover", 0) # 1: Allow branches to have a per branch failure route # 0: t_relay should fail if a branch fail modparam("tm", "failure_exec_mode", 1) # Allow locally generated 408 to be forgotten in case a reply is received. #modparam("tm", "faked_reply_prio", 2000) # USER location service route[LOCATION] { t_on_branch_failure("manage_branch_failure"); if (!lookup("location")) { .... } # Let's get back after 2.5s if no 100/Trying received. t_set_fr(0, 2500); # set_forward_no_connect(); #t_on_branch("SET_BRANCH_FAILURE"); route(RELAY); exit; } route[RELAY] { .... if (!t_relay()) { xlog("L_WARN","$rm|$ci|$fu|$tu|RELAY_FAILED"); sl_reply_error(); } exit(); } # set failure route per branch branch_route[SET_BRANCH_FAILURE] { xlog("L_WARN","$rm|$ci|$fu|$tu|SETTING_BrANCH_FAILURE_ROUTE"); t_on_branch_failure("manage_branch_failure"); } # manage failure routing cases per branch event_route[tm:branch-failure:manage_branch_failure] { xlog("L_WARN","$rm|$ci|$fu|$tu[$avp(dest_to)]|MANAGE_BRANCH_FAILURE"); }
_______________________________________________ SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list sr-users@lists.sip-router.org http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
-- Daniel-Constantin Mierla http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda Book: SIP Routing With Kamailio - http://www.asipto.com