Hi Andrew,<div><br></div><div>Thanks so much for this last message, that was key to my whole problem. I messed up network.yml so I reinstalled sipwise CE and only applied the machine's IP as domain, created trunk from FS and everything is fine, I can route calls incoming and outgoing now just fine.</div><div><br></div><div>I am wondering if you can help me with one more issue that has been causing me to nearly pull my hair out. My calls (both incoming and outgoing) are being dropped after 30-31 seconds every single time.</div><div><br></div><div>Naturally I am suspecting NAT issues, I have tried everything I can think of in FreeSwitch config with no changes, so I want to look at NGCP (as this is my in and out gateway) to see if the problem might be there.</div><div><br></div><div>One thing I noticed is that under "registere devices" for the FS trunk in the NGCP GUI, it shows "Nat=Yes" this should not be the case because both freeswitch and ngcp are on the 10.0.0.0/24 network.
Is there somewhere in the network configuration that I need to let NGCP know that FS (10.0.0.34) can connect to it without NAT because they are on the same subnet?</div><div><br></div><div>Thanks in advance!</div><div><br></div><div>Paul<br><br>On Sat, 13 Jul, 2013 at 3:54 AM, Andrew Pogrebennyk <apogrebennyk@sipwise.com> wrote:<br>
<blockquote type="cite">Hi,<br>
You should add <a href="http://10.0.0.40">10.0.0.40</a> as a domain in NGCP panel, if this is the address of your machine, and remove <a href="http://10.0.0.34">10.0.0.34</a> (this will delete subscribers provisioned in that domain too), because this is the FS address. After this there should be some noticeable progress.<br>
<br>
<br>
Paul <pasha@prosperity4ever.com> wrote:<br>
<br>
>On Fri, July 12, 2013 12:45 pm, Andrew Pogrebennyk wrote:<br>
>> Hi Paul,<br>
>><br>
>><br>
>> regarding the learning process, we have some diagrams to explain the<br>
>> system architecture here:<br>
><a href="http://sipwise.com/doc/2.8/spce/ar01s02.html">http://sipwise.com/doc/2.8/spce/ar01s02.html</a>,<br>
>> not sure if you've seen them. In short, every SIP message passes<br>
>through<br>
>> kamailio-lb which does basic DoS attack protection, and every INVITE<br>
>goes<br>
>> from lb to proxy them sems, which establishes a new SIP dialog with<br>
>_b2b-1<br>
>> suffix in Call-ID and sens INVITE to lb for passing it finally to<br>
>callee..<br>
><br>
>I will keep reading and rereading, thanks for the URL.<br>
><br>
>><br>
>><br>
>> On 07/13/2013 04:31 AM, Paul wrote:<br>
>><br>
>>> Andrew as always you are right :)<br>
>>><br>
>>><br>
>>> I adjusted a few more things in the trunk reg config, applied<br>
>settings<br>
>>> and now I have logs in the proxy log here they are:<br>
>><br>
>> Yes, I've noticed the 500 error in the lb log and was already<br>
>expecting<br>
>> to see the "Dropping local branch" message in proxy log:<br>
>><br>
>>> [...]<br>
>>> /usr/sbin/kamailio[11011]: INFO: <script>: Dropping local branch -<br>
>>> R=sip:gw+ngcp@10.0.0.34:5080;transport=udp;gw=ngcp<br>
>>> ID=C550303D@99.88.77.66<br>
>>><br>
>><br>
>> If PBX running on the same host by chance? Of have you added<br>
><a href="http://10.0.0.34">10.0.0.34</a><br>
>> as Domain in admin panel, same as <a href="http://10.0.0.40">10.0.0.40</a>? I'm asking because from<br>
>the<br>
>> log it looks like ngcp address is <a href="http://10.0.0.40">10.0.0.40</a>, the client is registered<br>
>> behind <a href="http://10.0.0.34">10.0.0.34</a>, but ngcp detects that address as local so refuses<br>
>to<br>
>> forward the call there.<br>
><br>
>PBX is not running on the same host.<br>
><br>
>NGCP: <a href="http://10.0.0.40">10.0.0.40</a><br>
>FreeSwitch: <a href="http://10.0.0.34">10.0.0.34</a> (where call should end up)<br>
><br>
>I did have <a href="http://10.0.0.34">10.0.0.34</a> as a domain in NGCP GUI, but not <a href="http://10.0.0.40">10.0.0.40</a>, just<br>
>added it, see if it makes any difference.<br>
><br>
>In domains I have --> <a href="http://10.0.0.34">10.0.0.34</a><br>
>In Subscribers I have --> DID@10.0.0.34<br>
><br>
>In freeswitch I setup a trunk that connects as follows:<br>
><br>
>user: DID@10.0.0.34<br>
>pass: password<br>
>realm: <a href="http://10.0.0.34">10.0.0.34</a><br>
>proxy: <a href="http://10.0.0.40">10.0.0.40</a><br>
><br>
>Both freeswitch and NGCP see the connection (freeswitch lists trunk as<br>
>registered, NGCP shows it under active devices for that subscriber).<br>
><br>
>Is there a link or document you can refer me to that explains how NGCP<br>
>analyzier from from-header to pull out the user the DID is attached to<br>
>and<br>
>then decides on how to forward it?<br>
><br>
>Thanks<br>
><br>
>Paul<br>
><br>
>><br>
>>><br>
>>> Based on this it looks to me like the Call-ID (error in logs) is<br>
>what<br>
>>> terminates the call and never actually pushes it out to the<br>
>freeswitch<br>
>>> trunk (I don't see anything in the freeswitch logs at all about a<br>
>call<br>
>>> coming in etc, and I do have a route for "public" context that will<br>
>>> catch all).<br>
>>><br>
>>> Thanks!<br>
>>><br>
>>><br>
>>> Paul<br>
>>><br>
>><br>
>><br>
<br>
-- <br>
Sent from my Android phone with K-9 Mail. Please excuse my brevity.<br>
-- <br>
Sent from my Android phone with K-9 Mail. Please excuse my brevity.</blockquote></div>