[asterisk-users] Re: Tracking the source of a disconnect?

[Tony Mountifield]

In article <4501787F.4070802 at asgardsrealm.net>,
Jamin W. Collins <jcollins at asgardsrealm.net> wrote:
> I have an asterisk box configured to perform media translation (TDM <-> 
> SIP).  With this configuration, calls are essentially only passing 
> through the asterisk box.  Thus, I would think that a disconnect request 
> should be received from one end of the call (SIP or TDM) into the 
> asterisk box as an end of the call.
> 
> However, periodically, I've been getting reports from users of being 
> disconnected in mid-conversation.  I've checked the system's logs for 
> any indication of problems and they all appear clean.  Eventually, I 
> enabled both PRI and SIP debugging in an effort to track down the 
> location of these disconnects.  At this time it appears that the 
> asterisk is initiating a disconnect of both the PRI and the SIP channel 
> (see the log snippet below).  However, there doesn't appear to be any 
> indication of why the asterisk is deciding to terminate the calls.
> 
> [...snip...]

It looks like the PRI connection is going down first, and when that channel
exits, it causes the SIP channel to be hung up. So concentrate on the PRI.

Try enabling intense PRI debugging "pri intense debug span N". You may want
to direct the PRI debugging to a file with "pri set debug file filename".

It's not clear from the log you posted whether q931_hangup() was called
because of a Q.931 message Asterisk received, or just because it decided to.
Hopefully, the intense debug would make that clear.

Cheers
Tony
-- 
Tony Mountifield
Work: tony at softins.co.uk - http://www.softins.co.uk
Play: tony at mountifield.org - http://tony.mountifield.org