[Asterisk-Users] Two FXOs getting bridged?

[Warren Burstein]

Dan Elder wrote:
> Hey all, I've been seeing this repeatedly and am wondering if anyone has a clue what's causing it.. at least once a day I see two zap fxo channels being bridged, and hanging..now, these two channels should never bridge, but they keep doing it.. any leads on where to look for what's causing it? here's what I see when I 'show channels'
>
> Zap/37-1             s at from-pstn:1        Up      Bridged Call(Zap/38-1)
> Zap/38-1             s at macro-dialout-trun Up      Dial(ZAP/g0/16509524400)
>
>
> 37 & 38 are the 1st two FXOs on the system, connected to a channel bank (CAC ABI) with POTS lines. When I spy on the channels, there is no activity (audio).... any leads as to where I should be looking to try to track this down? I'm unable to actually get an outbound trunk via the incoming lines, so I don't think it's a dialplan error..but who knows?.. using AAH 2.0 (* 1.2.1)
>   
Could someone have managed to transfer one outside call to another?  
That's happened a few times on my system, something like this: user 
calls first number, hangs up very briefly and picks up again (which is 
interpreted as a flash, so now the first outside call is on hold) calls 
the second number, eventually hangs up and now the two calls are 
bridged.  Here's how it looks on the console.  In this example I bridged 
a Zap FXS to a Sip phone but it looks the same with outside calls two 
Zap FXO's (I don't have any outside lines connected to my test system at 
the moment).  I don't know how to stop this from happening, either 
(without disabling call transfer, which I don't want to do) other than 
telling everyone to make sure that they don't flash when they really 
wanted to hang up.

I dialed 102, which dials Zap/2
    -- Starting simple switch on 'Zap/1-1'
    -- Executing Dial("Zap/1-1", "Zap/2") in new stack
    -- Called 2
    -- Zap/2-1 is ringing
    -- Zap/2-1 answered Zap/1-1
    -- Attempting native bridge of Zap/1-1 and Zap/2-1
here's where I flashed, putting Zap/2 on hold
    -- Starting simple switch on 'Zap/1-2'
    -- Started three way call on channel 1
    -- Started music on hold, class 'default', on channel 'Zap/2-1'
I dialed 139, which dials Sip/139
    -- Executing Dial("Zap/1-2", "Sip/139") in new stack
    -- Called 139
    -- SIP/139-9fde is ringing
    -- SIP/139-9fde answered Zap/1-2
here's where I hung up, bridging Zap/2 and Sip/139
    -- Stopped music on hold on Zap/2-1
    -- Hungup 'SIP/139-9fde<MASQ>'
  == Spawn extension (internal, 102, 1) exited non-zero on 'Zap/1-2
  == Spawn extension (internal, 139, 1) exited non-zero on 'Zap/1-2'
    -- Hungup 'Zap/1-2'
badger*CLI> show channels
Channel              Location             State   
Application(Data)            
Zap/2-1              139 at internal:1         Up      Bridged 
Call(SIP/139-9fde)   
SIP/139-9fde         102 at internal:1       Up      
Dial(Zap/2)               
2 active channels
1 active call

The "show channels" looks like yours, which is why I think it might be 
the same thing.  Look in the log file if the lines that show how this 
happened have scrolled off your screen.

One additional mystery is that I don't know why these calls persist.  
When I hang up either of the bridged extension on my test system, the 
bridged call ends.  When a single outside call is hung up on the other 
side, asterisk notices.  I don't have enough phone lines and cellphones 
to test if this works when two outside lines are bridged.  Does external 
hangup detection work on your system?