November 2018 - FreeSWITCH-users - FreeSWITCH Mailing Lists

Access to specific B leg headers on failed bridge
by David Witham 28 Nov '18

28 Nov '18

I have a problem where an upstream provider will correctly send a SIP 486 Busy Here from their voice gateway with a Reason header containing Q.850 cause=17 when a B party is genuinely busy. However if we exceed certain traffic thresholds, instead of receiving a SIP 503 with a retryable cause code, their proxy sitting in front of their voice switch sends us a SIP 486 Busy Here and no reason header. Different user agent and also contains a Server: header that the voice gateway responses do not. Therefore we start failing calls that we would otherwise route advance to another upstream. I want to be able to discern the difference between the two responses and route advance on the false user busy. I've tried using variations of the example LUA code found here https://freeswitch.org/confluence/display/FREESWITCH/Cause+Code+Substitutio… but what I have found is that, of the variables I seem to have access to at this point in the call flow, sip_invite_failure_status=486 and last_bridge_hangup_cause=USER_BUSY are set the same in both cases. I think I need to be able to check for the absence of a Reason header or presence of the Server header to know that I have a false user busy. My question is then, is it possible to obtain the specific headers that are present in the B leg response after a failed bridge? I've tried addressing sip_h*, sip_bye_h* type variables but none seem to be populated. I'm guessing that I'm trying to address them in the A leg but they're only present in the B leg. Any tips or advice would be greatly appreciated. David

1 0

Problem with German Telekom SIP Trunk
by ITutu - Mailings 27 Nov '18

27 Nov '18

Hi, I’ve setup FreeSWITCH (using FusionPBX) to connect to the German Telekom using their SIP-Trunk product. Basically everything works fine except I have to restart the external SIP Profile every 1-2 weeks because of a 403 FORBIDDEN message returned by the Telekom server. After some investigation I think I know why this happens, but I have no idea how to solve it. The DNS SRV records from the Telekom return 3 servers with different priorities. - You can check that with: nslookup -querytype=SRV _sip._tcp.reg.sip-trunk.telekom.de FreeSWITCH takes the one with the highest prio and registers. - e.g. serverA - Also every 240 seconds a re-registration happens Then everything works as expected - All INVITES go to serverA as well (because reg.sip-trunk.telekom.de is also the outbound proxy) Once in a while the Telekom changes the prio of the servers - So the next INVITE goes to serverB - And this is where I get the 403 because they expect the INVITE to be sent to the server FreeSWITCH registered before (in this example serverA) So I see two problems here I’d like to solve: When re-registering every 240 seconds it looks like FreeSWITCH doesn’t do another DNS lookup and doesn’t re-register to the server with highest prio - How can I force FreeSWITCH to do this DNS lookup before every registration and choose the new “best” server? - How can I force FreeSWITCH to send the INVITES to the server it registered before (and not to the one with the currently highest prio)? When doing some research I found the 3CX has some new option implemented especially for this problem: “Force invites to be send to IP of registrar” Please see: https://www.3cx.de/blog/deutsche-telekom-sip-trunk-wichtig/ Is something like this also possible with FreeSWITCH? Thanks, Klaus

2 2

sip capture
by Markus Bönke 25 Nov '18

25 Nov '18

Hello, is sip-capture supported in 1.8.2? I’ve enabled sip capturing (as recommended at https://github.com/sipcapture/homer/wiki/Examples%3A-FreeSwitch <https://github.com/sipcapture/homer/wiki/Examples:-FreeSwitch>) <param name="sip-capture" value="yes"/> <param name="capture-server" value="udp:192.168.156.57:9060;hep=3;capture_id=2008“/> but in the network trace I cannot see the any HEP packets are send to the capture server. Do I need to configure something else? Thanks and regards Markus

4 11

Call recording and intercept interplays
by Ryan Delgrosso 24 Nov '18

24 Nov '18

Greetings all, I am looking to trigger session recording when the A leg of a call is actually bridged to an endpoint. I have been successful when using execute_on_answer or when using bridge_pre_execute_aleg when the call is answerd by the original endpoint, but when the original A leg is intercepted with the intercept application I'm not able to get any of these triggers to fire as desired. I don't want to simply begin recording without a bridge as I then just end up recording the call unnecessarily if it goes to voicemail. Can anybody give me some guidance on what callbacks are fired on a leg when it is intercepted and bridged to the intercepting leg or perhaps another way to accomplish the goal of only recording upon bridge. Thanks in advance -Ryan

1 0

mod_verto dialplan variables
by Nuno Bessa 24 Nov '18

24 Nov '18

Hi all, I'm using mod_verto with the javascript library and it's working great except for one thing I haven't quite figured out yet. 95% of our calls are outbound, from mod_verto to PSTN and my outbound dialplan is set to record all sessions. Before mod_verto, we used a desktop app developed by me, wich in turn received a variable from the dialplan on A-leg set like this: <action application="set" data="sip_rh_X-RecordingFilename=${strftime(%Y)}/${strftime(%m)}/${strftime(%d)}/${strftime(%H)}/${strftime(%M)}/${destination_number}_${caller_id_number}.wav"/> My application would store that variable along with all client details, for later supervision. Recordings are stored using YYYY/MM/DD/HH/MM/LEG-B-NUMBER_LEG-A-EXTENSION.WAV format, and I need to acess this variable, because client's clock may slightly differ from freeswitch server, and the path to the recording would be wrong when stored in MySQL. With mod_verto and the javascript library, I can't read custom variables and get them on the javascript dialog object, more specifically on dialog.params, which is the place to read them I believe. This is my complete dialplan: <include> <extension name="Call_Verto"> <condition field="source" expression="mod_verto"> <condition field="destination_number" expression="(^\d{9,14}$)"> <action application="set" data="absolute_codec_string=^^:PCMA:PCMU"/> <action application="set" data="RECORD_DATE=${strftime(%Y-%m-%d %H:%M)}"/> <action application="set" data="RECORD_STEREO=false"/> <action application="record_session" data="$${basedir}/Program Files/FreeSWITCH/recordings/archive/outbound/${strftime(%Y)}/${strftime(%m)}/${strftime(%d)}/${strftime(%H)}/${strftime(%M)}/${destination_number}_${caller_id_number}.wav"/> <action application="set" data="verto_h_recording_filename=${strftime(%Y)}/${strftime(%m)}/${strftime(%d)}/${strftime(%H)}/${strftime(%M)}/${destination_number}_${caller_id_number}.wav"/> <action application="bridge" data="sofia/gateway/IPPBX/$1"/> </condition> </condition> </extension> </include> How can I set the verto_h_recording_filename variable in order to acess it with verto javascript client being the A-leg a mod_verto client? I searched all over but couldn't find a solution to my specific case, and documention is still a little sparse unfortunatelly. Thanks in advance for your help!

3 3

T.38 UDP port range
by Kevin Olbrich 23 Nov '18

23 Nov '18

Hi! Which ports do I need to open to allow T.38? The RTP range can be set but I was unable to find such a setting for T.38. Asterisk for example needs range 4000:4999. Kind regards Kevin

3 2

NAT Problem with UPNP only for RTP - SIP tcp keept open
by Paul Muaddib 21 Nov '18

21 Nov '18

Hi at the moment I have the setup Internal.xml ext-rtp-ip = auto-nat ext-sip-ip = auto-nat External.xml ext-rtp-ip = “my_local_freeswitch_ip” ext-sip-ip = auto-nat RTP ports (only) are begin opened by UPNP and SIP via TCP is being keep open with “expire-seconds=600” and “register=true” (Is there an alternative for the expire-seconds ? nat-options-ping is only for endpoints registering to freeswitch, right?) The reason why I am doing this is, that I dont want to open up a SIP Port in the Firewall. But this setup up doesn’t seem right. Audio is working, inbound and outbound, both ways. But the echo application is not. When I change External.xml ext-rtp-ip = “auto-nat” I get one-way audio for inbound calls for the callee. The caller can be heard. I tried to analyses it in wireshark, but I don’t understand it. The sip provider keeps sending INVITE messages while freeswitch is confirming it with OK but keeps going on until the connection breaks. Thank you for your help

2 2

Generate uniqueid like asterisk
by Mickael Hubert 21 Nov '18

21 Nov '18

Hi all, currently, I use asterisk uniqueid to lot of internals processes Ex: 001-1542296620.1319230 I want to change asterisk by FS, but I have to keep this uniqueid format. I know the format of the uniqueid : [server id]-[timestamp].[random integer] How can I generate it from FS please ? I already seen with mod_dialplan_asterisk module, but I'm not sure if can I use only UNIQUEID variable (I'm not need the asterisk dialplan) Or maybe use other variables Ex: unixtimestamp and create random interger... thanks in advance ++

7 10

scratch sound at beginning of SRTP call
by Maarten Ureel 20 Nov '18

20 Nov '18

Hello I am running a setup where calls from the outside are not encrypted with SRTP (from our SIP trunk), and locally (between phone & FS) they are. I put in the vars.xml: <X-PRE-PROCESS cmd="set" data="rtp_secure_media=optional:AES_CM_128_HMAC_SHA1_80"/> This all works fine, however with one disturbing fact; when you pick up the call you hear a loud "static" noise. I think this is the phase where the suite is negotiated, because it's the same as when we force SRTP on FreeSWITCH and disable it in the phone (but then the call fails afterwards of course). So what I would need to accomplish: * Either the phone needs to negotiate the SRTP before it's answered (is that even possible?), e.g. during ring * If the negotiation has to take place after answering, that it doesn't generate this loud noise in the meantime. It could also be due to the used phone (Yealink, T42G and T46G). With RTX phones we didn't have this issue yet but I only tested it with one device. But if someone has some pointers on what causes this... Thanks for your input! Regards Maarten

1 0

Monitoring mod_odbc_cdr
by morfair＠gmail.com 19 Nov '18

19 Nov '18

Hello all! I use mod_odbc_cdr for store CDR in MySQL for billing. Today I saw that no records in DB in 2 weeks. IT IS FAIL!! I went to fs_cli and type `reload mod_odbc_cdr`. After that all work again. But I lost records in two weeks. How to monitor mod_odbc_cdr and prevent risks with its?

6 5