[Asterisk-Users] specify maximum call duration

Philipp von Klitzing klitzing at pool.informatik.rwth-aachen.de
Mon Dec 29 08:50:36 MST 2003


Hi Steven!

> For your extended part of the original question.
> [incoming]
> exten => _.*,1,answer
> ;exten => _.*,2,agi(timeout-lookup.agi) ; alternative
> exten => _.*/some match,2,Absolutetimeout(360)
> exten => _.*,2,noop
> exten => _.*,3,goto(realcontext,${EXTEN},1)
> 
> Does this look hard or tedious? 
> 
> > I _do_ think it would make sense to be able to define such a timeout for 
> > a specific peer, or even as a global varaible valid for the entire * 
> > server.
> 
> Maybe you don't yet grasp how easy things can already be done. Only
> thing I think might be nice, but I don't really see a use for yet would
> be a method to define extra variables on any of the VoIP user
> definitions. 

Your suggestion looks fine, however it only covers incoming calls for 
that specifc context. So we'll need to care of outgoing calls as well - 
and if you have a complex dialplan you will most likely not have just two 
places where this can be set, e.g. you'll need to deal with internal 
calls as well etc.

Let me explain why I joined this discussion: With a non-trivial 
extensions.conf you can easily screw up things, I guess it has happened 
to most of us here one way or another (forgot a , or misspelled a 
keyword, or ran copy-paste action and forgot to adjust the extension 
number etc). Likewise there is a good chance to miss out on a possible 
path a call could take with all those t and s and exiting apps and so 
forth. 

For example, I am concerned about an international outgoing ISDN call 
hanging for whatever reason, and I would feel MUCH more safer if I had a 
global setting that would catch the call even if I goofed in 
extensions.conf. Of course the most common problem are stuck SIP calls. 
By now I learned to reload and check the log immediately after any change 
to extensions.conf, but that still can't protect you from every possible 
(human) error...

Cheers, Philipp





More information about the asterisk-users mailing list