[asterisk-users] Problem with libpri / asterisk

[Nicolas Ross]

Hi all !

We currently have an asterisk box that is rather old (runs Asterisk
1.4.21.2), and it's connected to the PSTN with a sangoma A104d card.

Now we have a new PRI at another location, and I use that occasion to
build 2 new servers, one to replace our aging one and a new one for this new
pri.

So I downloaded the lastest libpri / asterisk / wanpipe driver, but the
previous version of dahdi (2.5), since the latest wanpipe isn't compatible
with dahdi 2.6. All is built from source

Now, all seems to be working OK. I can connect a SIP phone to my new box,
make calls to the outside, receive calls etc.

But, I can't seem to bridge a call. So on my new server, with the new PRI, I
got a Sangoma a104 card (no echo-canceler on this one).

In my extensions.ael, I got this :

418nxxxxx1 => {
  Answer();
  Wait (2);
  Playback(demo-thanks);
  Dial(${TRUNK}/418nxxxxx2);
};

TRUNK is DAHDI/G1

Where 418nxxxxx1 is a DID on my new PRI and 418nxxxxx2 is my cellphone
number.

When I do a call from my home phone or cell phone to my new PRI to
418nxxxxx1, I hear the demo-thanks file, and then it dials out. My cellphone
rings, but as soon as I pick up the call, the calls hangs up :

    -- Accepting call from '418nxxxxx2' to '418nxxxxx1' on channel 0/1, span 
1
    -- Executing [418nxxxxx1 at ael-default:1] Answer("DAHDI/i1/418nxxxxx2-b", 
"") in new stack
    -- Executing [418nxxxxx1 at ael-default:2] Wait("DAHDI/i1/418nxxxxx2-b", 
"2") in new stack
    -- Executing [418nxxxxx1 at ael-default:3] 
Playback("DAHDI/i1/418nxxxxx2-b", "demo-thanks") in new stack
    -- <DAHDI/i1/418nxxxxx2-b> Playing 'demo-thanks.ulaw' (language 'fr')
    -- Executing [418nxxxxx1 at ael-default:4] Dial("DAHDI/i1/418nxxxxx2-b", 
"DAHDI/G1/418nxxxxx2") in new stack
    -- Requested transfer capability: 0x00 - SPEECH
    -- Called DAHDI/G1/418nxxxxx2
    -- DAHDI/i1/418nxxxxx2-c is proceeding passing it to 
DAHDI/i1/418nxxxxx2-b
    -- DAHDI/i1/418nxxxxx2-c is ringing
    -- DAHDI/i1/418nxxxxx2-c is making progress passing it to 
DAHDI/i1/418nxxxxx2-b
    -- DAHDI/i1/418nxxxxx2-c answered DAHDI/i1/418nxxxxx2-b
    -- Native bridging DAHDI/i1/418nxxxxx2-b and DAHDI/i1/418nxxxxx2-c
    -- Span 1: Channel 0/1 got hangup request, cause 16
    -- Hungup 'DAHDI/i1/418nxxxxx2-c'
  == Spawn extension (ael-default, 418nxxxxx1, 4) exited non-zero on 
'DAHDI/i1/418nxxxxx2-b'
    -- Hungup 'DAHDI/i1/418nxxxxx2-b'

BUT, if I originate the call from my curent PRI, it goes in and out and all
is well. I noticed that if the calls go trough correctly and hangup 
manually, it also stats the exact same thing (cause 16). So the above 
console output might not be that much usefull...

I've had a case open with Sangoma for this issue, and they suggested I go 
the libpri/asterisk for more help debuging this issue, since on their end, 
the disconnect comes from the telco...

They suggested I try a different version of asterisk, wich I did to no 
avail, or try there NBE product instead of libpri...

So, did anybody ever encontered something like that ? What steps should I 
take to diagnose the problem furhter ?

Thanks for any help.