Re: Timming Source

Hello,

No, Yate doesn't need a timing source at all.

Diana

Pablo Prats wrote:
> Hi,
>
> Does Yate needs a timing source like asterisk??
>
> If yes. Is sangoma USB device enough?
>
> thanks!
>
> Pablo

Re: IAX2 Trunking / Yate Solution

Hello,

Just start more than one Yate on the same server. The limit it's at 
around 300 calls, so if you have a Yate for every 200 calls you should 
be safe. You can use SIP as a protocol between Yate's.

Diana

Pablo Prats wrote:
> Hi,
>
> I'm new to YATE.
>
> I leave out asterisk for this one, because it makes more sense YATE 
> for this project.
>
> I need to use IAX2 protocol for a large application, 2,000 concurrent 
> calls/g729 pass-through for example. I know that IAX2 don't scale well 
> because it only have one thread to receive datagrams for all the 
> calls. I need a work around this problem because I really need IAX2 to 
> save as much bandwidth as I can.
>
> In Mexico the bandwidth is very expensive.
>
> My set-up is like this:
> Note: myClient is the call originator.
>
> myClient<-----SIP----> myServer1(USA) <-----IAX2/T----> 
> myServer2(MX)<---SIP---> quintum <--E1--> PSTN (Termination)
>

Callfork and libyate.php

ThreadedSource API change notification

Hi all!

SVN Rev. 2827 (2009-09-07) brings changes in how ThreadedSource operates and 
the exposed API. This will fix some subtle crashes happening very 
infrequently when prompts are quickly replaced and/or channels are destroyed 
in an IVR.

Additionally there are some related changes in the modules that use this class 
(tonegen, wavefile, moh, extmodule, alsachan, osschan) as well as some fixes 
and cleanups, especially in the wavechan and alsachan modules.

Only if you are a developer and you created your own channels that use the 
ThreadedSource class you should pay attention below.

Short:
- A ThreadedSource object is kept referenced by its internal thread.
- Such an object will not be promptly destroyed when removed from the last 
DataEndpoint (even before it was possible to have other references).
- The ThreadedSource::looping() will return false when it's time to exit from 
the run() method, that is:
	- The helper thread is asked to terminate by calling stop()
	- The helper thread was replaced or looping() is not called from that thread
	- The helper thread holds the last reference to the source
	- The engine is exiting
- Use the new clearData() methods of the DataEndpoint and CallEndpoint classes 
to make sure a source or consumer was removed from its owner. These take care 
of the condition where the source or consumer was replaced juuust before 
trying to clear it out.
- The DataNode::attached() method provides a chance to test the current state 
when the source or consumer is inserted or removed from a DataEndpoint.

All best,

Paul

Re: Callfork and libyate.php

Hi,

The IAX target should be

iax/user <at> host/called_number

Marian

Alejandro Ruiz wrote:
> I am trying to use the callfork module, with regexroute and through a php
> script.
> everything worked fine while I used h323 and sip (with no registration),
> trying to use iax is quite frustrating...
> eventually I will need to do some fallback with two iax peers.
> 
> I am using without callfork somthing like:
> iax/user:pass <at> 1.1.1.1<user%3Apass <at> 1.1.1.1>
> ;called=123456789
> 
> the call fork: fork iax/user:pass <at> 1.1.1.1
> <user%3Apass <at> 1.1.1.1>;called=123456789
> | iax/user:pass <at> 2.2.2.2 <user%3Apass <at> 2.2.2.2>;called=222222
> 
> wont work, something similar happens in the php script, whee I have to put
> the destination called number, as globally called whith something like:
> "$ev-params["called"] = 22222"
> 
> any ideas or workaround will be apreciated.
> 
> thanks
> alejandro
> 

Re: ThreadedSource API change notification

... and remember that now you need to call ThreadedSource::cleanup() if you 
reimplement it an descendent class.

See Rev. 2829 

Re: Callfork and libyate.php