28 Jul
2011
28 Jul
'11
5:57 p.m.
Hi Daniel, Hi Klaus,
after hours of searching I think we found the problem. This is what our
route that actually sends out the packets looks like:
route[2] {
xlog("........................................... START OF ROUTE
2....................................\n");
xlog("AVP LEG_ID has Value $(avp(s:leg_id))\n");
$(avp(s:dsttype)) = 0;
if ($rd =~ "pstnout") { $(avp(s:dsttype)) = 2; }
if ($rd =~ "media") { $(avp(s:dsttype)) = 1; }
if (isflagset(0)) {
$(avp(s:dsttype)) = 3;
t_on_reply("4");
$uac_req(method)="PDDTRACKING";
$uac_req(ruri)="sip:store@172.20.21.3:5160";
$uac_req(furi)="sip:kamailio@sipgate.net";
$uac_req(hdrs)="Content-Type: text/timetracking-csv\r\n";
pv_printf("$uac_req(body)",
"$ci,$fU,$tU,$TV(Sn),$rm,$rd");
uac_req_send();
}
if ((!defined ($(avp(s:tariffannouncement))) ||
!($(avp(s:tariffannouncement)) == "true")) && method=="INVITE"
&& defined
$(avp(s:cid)) && !($(avp(s:cid)) == "")) {
$uac_req(method)="TIMETRACKING";
$uac_req(ruri)="sip:store@172.20.21.3:5160";
$uac_req(furi)="sip:kamailio@sipgate.net";
$uac_req(hdrs)="Content-Type: text/timetracking-csv\r\n";
pv_printf("$uac_req(body)",
"$(avp(s:cid)),$ci,$fU,$tU,$TV(Sn),$(avp(s:dsttype)),$rm,$rd");
uac_req_send();
}
xlog("........................................... AFTER UAC REQUEST
....................................\n");
xlog("AVP LEG_ID has Value $(avp(s:leg_id))\n");
t_on_reply("ABC");
if (!t_relay()) {
xlog("L_ERR", "route(2): error to <$tu> from
<$fu>\n");
if (method=="INVITE" || method=="ACK") {
unforce_rtp_proxy();
}
sl_reply_error();
return;
};
return;
}
I probably need to explain that a little. We have another kamailio process
just for accounting. We do it separately for being independent of the
database in our "routing Kamailio", because there would be the chance of
blocked processes, if the database is too slow, if we did it in the routing
Kamailio.
We try to to measure Post Dial Delay when sending calls out to the carrier,
and we monitor the time differences for one call each time Kamaiio is
passed. That's all the other process does. So we send handcrafted SIP
packages with the needed information with uac_req_send(). And it looks like
the AVPs get lost at that point.
That's what the log says:
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17684]: ERROR: <script>:
........................................... START OF ROUTE
2....................................
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17684]: ERROR: <script>: AVP LEG_ID
has Value 9962182
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17684]: ERROR: <script>:
........................................... AFTER UAC REQUEST
....................................
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17684]: ERROR: <script>: AVP LEG_ID
has Value <null>
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17681]: NOTICE: <script>: REPLY
START
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17681]: NOTICE: <script>: reply: 100
Trying - F=sip:anonymous@sipgate.de T=sip:0211XXXXXXXX@sipgate.de SRCIP=
217.116.120.220:5060
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17681]: NOTICE: <script>: Now in
named reply route...............................................
Jul 28 16:41:35 hagi /usr/sbin/kamailio[17681]: ERROR: <script>: AVP LEG_ID
has Value <null>
If I disable the UAC part of route[2], I get the following output:
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17404]: ERROR: <script>:
........................................... START OF ROUTE
2....................................
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17404]: ERROR: <script>: AVP LEG_ID
has Value 5770077
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17404]: ERROR: <script>:
........................................... AFTER UAC REQUEST
....................................
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17404]: ERROR: <script>: AVP LEG_ID
has Value 5770077
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17403]: NOTICE: <script>: REPLY
START
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17403]: NOTICE: <script>: reply: 100
Trying - F=sip:anonymous@sipgate.de T=sip:0211XXXXXXXXX@sipgate.de SRCIP=
217.116.120.220:5060
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17403]: NOTICE: <script>: Now in
named reply route...............................................
Jul 28 16:25:20 hagi /usr/sbin/kamailio[17403]: ERROR: <script>: AVP LEG_ID
has Value 5770077
The behavior was different in Kamailio 1.5, and I think using the uac module
shouldn't destroy AVPs.
Am I missing something?
Best regards,
Sebastian
P.S.: Of course, the named onreply_route is executed AFTER the standard
onreply_route. I misinterpreted my logs last week.
On Fri, Jul 15, 2011 at 2:32 PM, Daniel-Constantin Mierla <miconda(a)gmail.com
wrote:
Hi Sebastian,
On 7/15/11 11:42 AM, Sebastian Damm wrote:
--
Dipl.-Inf.
Sebastian Damm - VoIP-Engineer - damm(a)sipgate.de
sipgate GmbH - Gladbacher Str. 74 - 40219 Düsseldorf
HRB Düsseldorf 39841 - Geschäftsführer: Thilo Salmon, Tim Mois
Steuernummer: 106/5724/7147, Umsatzsteuer-ID: DE219349391
www.sipgate.de - www.sipgate.at - www.sipgate.co.uk
Hi,
On Fri, Jul 15, 2011 at 9:02 AM, Klaus Darilion
<klaus.mailinglists(a)pernau.at> wrote:
I am a bit confused about what you explain above, with "named route is
executed before the default reply_route". Can you send like the structure of
the config for this case? I mean the routes involved and the calls of
t_on_reply(). Is it like for example:
route {
...
$avp(xyz) = 1;
t_on_reply("ABC");
t_relay();
exit;
}
onreply_route[ABC] {
...
xlog("avp(xyz) is $avp(xyz)\n");
...
}
Cheers,
Daniel
--
Daniel-Constantin Mierla -- http://www.asipto.com
Kamailio Advanced Training, Oct 10-13, Berlin: http://asipto.com/u/kat
http://linkedin.com/in/miconda -- http://twitter.com/miconda
The named reply routes are only executed if
t_on_reply() was called for
the request. This reply route will be executed after the default
reply-route. It is triggered by tm module.
I just inserted a new reply_route, which just prints out some
variables (and does avp_print). From what I see in the logs now, the
named route is executed before the default reply_route. And there are
no AVPs in the named route. Actually, I don't even need the AVPs in
the replies, I need them to be there, when the 200 OK comes in or in
failure route when the call is cancelled. In both cases the AVPs are
<null> if i directly address them.
Any more ideas?