Liviu, Yes, I meant it in the more general sense of re-sending the request, but of course one must be mindful of wording when these terms have specific meanings as this one does! It would indeed be a new transaction. Thanks for catching that!
Ben Newlin From: Users <[email protected]> on behalf of Liviu Chircu <[email protected]> Reply-To: OpenSIPS users mailling list <[email protected]> Date: Tuesday, July 17, 2018 at 9:03 AM To: "[email protected]" <[email protected]> Subject: Re: [OpenSIPS-Users] Multiple branches for a single AOR Hi Ben, Just for the sake of completeness, if we do decide to retry a request within a failure route, we will not do a "retransmission", but actually create a brand new transaction (with a unique ";branch=" param) on the SIP proxy's UAC side (the outgoing side). Best regards, Liviu Chircu OpenSIPS Developer http://www.opensips-solutions.com On 17.07.2018 15:54, Ben Newlin wrote: Gerwin, Specifically, it is the t_relay() call within your failure_route. t_relay() is used to send requests, not responses. The automatic action of failure_route if no retransmission is attempted is to send the reply back upstream [1]. So you don’t have to do anything if that is the functionality you desire. [1] http://www.opensips.org/Documentation/Script-Routes-2-4#toc3 Ben Newlin From: Users <[email protected]><mailto:[email protected]> on behalf of Liviu Chircu <[email protected]><mailto:[email protected]> Reply-To: OpenSIPS users mailling list <[email protected]><mailto:[email protected]> Date: Tuesday, July 17, 2018 at 8:50 AM To: "[email protected]"<mailto:[email protected]> <[email protected]><mailto:[email protected]> Subject: Re: [OpenSIPS-Users] Multiple branches for a single AOR Hi Gerwin, Inside your failure route, you are always attempting a retry of any failed request. This logic conflicts with your initial statement that "The intent is to ensure that the response code gets sent through to A-party". Best regards, Liviu Chircu OpenSIPS Developer http://www.opensips-solutions.com On 17.07.2018 14:35, Gerwin van de Steeg wrote: Folks, I'm trying to narrow down a 482 Merged Request problem on calls from one SIP device to another via OpenSIPS 2.4.1. Yealink T41P SIP device (A-party), calls via OpenSIPS, to another AOR owned by a Zoiper5 device (B-party). The intent is to ensure that when the B-party rejects the call with a 486 Busy Here, that the response code gets sent through to A-party. However what I'm seeing is the 486 gets sent to OpenSIPS which ACK's it, but doesn't go anywhere from there, and then something causes a second invite to be sent from OpenSIPS to the B-party which then responds of course with 482 Merged Request. The call as it is progressing through the call flow seems to be starting a second branch to the AOR (only one SIP device registered using UDP per AOR). What would be causing that second call so that I can eliminate it and get to the behaviour I'm expecting. Just using a slightly modified residential default config template with websocket support (the problem was noticed using SIP.JS but exists also in generic SIP device to SIP device calls). Image containing sngrep of call: https://imgur.com/RCZXkO6 Subscribers are in the form of <username>@<domain> With an alias setup for an extension number. ie. alfred.anderson@... = 552 alice.bell@... = 553 excerpt from opensips.cfg if ($rU==NULL) { # request with no Username in RURI send_reply("484","Address Incomplete"); exit; } $acc_extra(src_ip) = $si; # source IP of the request $acc_leg(caller) = $fu; $acc_leg(callee) = $ru; # apply DB based aliases if (alias_db_lookup("dbaliases")) { xlog("Alias lookup success [$fu/$tu/$ru/$ci]"); } else { xlog("Alias lookup failure [$fu/$tu/$ru/$ci]"); } # do blind callforward lookup if (avp_db_load("rU", "$avp(callfwd)")) { t_reply("181", "Call Is Being Forwarded"); $ru = $avp(callfwd); xlog("forwarded call to: $avp(callfwd)"); route(relay); exit; } # apply transformations from dialplan table dp_translate("0", "$rU/$rU"); # check if the call needs to be routed to freeswitch route(freeswitch); # here we would set the redirect URI if it had one route(lookup); } route[lookup] { script_trace(1, "$rm from $si, rur=$ru", "me"); xlog("route:lookup"); # do lookup with method filtering if (!lookup("location","m")) { xlog("lookup failure"); t_newtran(); if (!db_does_uri_exist()) { xlog("$cfg_line: URI doesn't exist"); send_reply("420", "Bad Extension"); exit; } t_reply("404", "Not Found"); exit; } # when routing via usrloc, log the missed calls also do_accounting("db","missed"); route(relay); } route[freeswitch] { xlog("route:freeswitch"); if (!is_method("INVITE")) { return; } # if the called number begins with the right dialplan redirect it to freeswitch # here we take everythign prefixed with a *, strip it, and send it to freeswitch if ($rU=~"^\*") { strip(1); $du = "sip:10.23.4.192:50600<http://10.23.4.192:50600>"; route(relay); } } route[relay] { xlog("route:relay: Relaying: method=$rm"); # for INVITEs enable some additional helper routes if (is_method("INVITE")) { t_on_branch("per_branch_ops"); t_on_reply("handle_nat"); t_on_failure("missed_call"); } else if (is_method("BYE|CANCEL")) { # cancel the rtpengine transcoding rtpengine_delete(); } if (!t_relay()) { send_reply("500","Internal Error"); } exit; } branch_route[per_branch_ops] { script_trace(1, "$rm from $si, rur=$ru", "me"); xlog("[$ci/$T_branch_idx] branch_route:per_branch_ops: new branch at $ru\n"); # WebSocket specific handling with NORMAL SDP negotiation # assumes SDP offer in the INVITE from the UAC, and SDP # answer is in 200 OK from the UAS if (!is_method("INVITE") || !has_body("application/sdp")) return; if (isflagset(SRC_WS) && isbflagset(DST_WS)) $var(rtpengine_flags) = "ICE=force-relay DTLS=passive"; else if (isflagset(SRC_WS) && !isbflagset(DST_WS)) $var(rtpengine_flags) = "RTP/AVP replace-session-connection replace-origin ICE=remove"; else if (!isflagset(SRC_WS) && isbflagset(DST_WS)) $var(rtpengine_flags) = "UDP/TLS/RTP/SAVPF ICE=force"; else if (!isflagset(SRC_WS) && !isbflagset(DST_WS)) $var(rtpengine_flags) = "RTP/AVP replace-session-connection replace-origin ICE=remove"; # only enable transcoding if websocket call for now if (isflagset(SRC_WS) || isbflagset(DST_WS)) { rtpengine_offer("$var(rtpengine_flags)"); } } onreply_route[handle_nat] { script_trace(1, "$rm from $si, rur=$ru", "me"); xlog("[$ci/$T_branch_idx] onreply_route:handle_nat: $ru\n"); # WebSocket specific handling with NORMAL SDP negotiation # assumes SDP offer in the INVITE from the UAC, and SDP # answer is in 200 OK from the UAS if (!has_body("application/sdp")) return; if (isflagset(SRC_WS) && isbflagset(DST_WS)) $var(rtpengine_flags) = "ICE=force-relay DTLS=passive"; else if (isflagset(SRC_WS) && !isbflagset(DST_WS)) $var(rtpengine_flags) = "UDP/TLS/RTP/SAVPF ICE=force"; else if (!isflagset(SRC_WS) && isbflagset(DST_WS)) $var(rtpengine_flags) = "RTP/AVP replace-session-connection replace-origin ICE=remove"; else if (!isflagset(SRC_WS) && !isbflagset(DST_WS)) $var(rtpengine_flags) = "RTP/AVP replace-session-connection replace-origin ICE=remove"; # only enable transcoding if websocket call for now if (isflagset(SRC_WS) || isbflagset(DST_WS)) { rtpengine_answer("$var(rtpengine_flags)"); } } failure_route[missed_call] { script_trace(1, "$rm from $si, rur=$ru", "me"); xlog("[$ci/$T_branch_idx] failure_route:missed_call: incoming failure response to $rm <- $T_reply_code/$T_ruri"); if (t_was_cancelled()) { xlog("[$ci/$T_branch_idx] was cancelled"); exit; } do_accounting("db", "missed"); if (!t_relay()) { send_reply("500","Internal Error"); } else { xlog("[$ci/$T_branch_idx] Relay success $rm/$T_reply_code"); } } Cheers, Gerwin _______________________________________________ Users mailing list [email protected]<mailto:[email protected]> http://lists.opensips.org/cgi-bin/mailman/listinfo/users _______________________________________________ Users mailing list [email protected]<mailto:[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
