[Asterisk-Dev] chan_exosip2

[Olle E. Johansson]

Kevin Hanson wrote:
> Olle E. Johansson wrote:
> 
>>> I have a customer with a receptionist that has a Polycom 601 w/ side
>>> car.  They are running Asterisk 1.2 rc 2.  I have all the other
>>> extensions setup as buddies on the receptionist phone.  When the phone
>>> first boots up it appears that presence is working great.  An associate
>>> is on the phone and the associated button and light on the receptionist
>>> phone indicate "in use".
>>>   
>>
>> Use "sip show inuse" to track the inuse counter. There may be errors in
>> this logic.
>>  
>>
> I am getting inconsistent results with this.  I had two sip channels
> with call-limit=5.  'sip show inuse' showed them correctly.  I removed
> call-limit from one of the channels, reloaded and 'sip show inuse'
> showed that the channel had been removed from the user section, but not
> peer (or vice versa, can't remember).
Yes, the peer is *not* reloaded if it's active (registred).

> 
>>> However, after a period of time presence stops working, or only works on
>>> some of the subscribed extensions (it's not consistent which ones are
>>> working at any particular time).  [One note...the receptionist phone
>>> always correctly shows that the other extensions are "on line" (i.e.
>>> registered w/ asterisk), so "on line" state works.]  It just doesn't
>>> always correctly show the other device states I need...ringing or in
>>> use.  As a work-around they are using Flash Operator Panel to see device
>>> state.
>>>   
>>
>>
>> Again, please check with "sip show inuse"
>>  
>>
> I'm also getting weird results with 'show hints'.  In my lab I have one
> buddy on the polycom watching another phone.  'show hints' shows 1
> watcher.  Great.  The phone shows when the monitored extension is 'in use'.
> 
> During my reload above while testing 'sip show inuse', watchers goes to
> zero.  So, I disable / renable buddy watch on the polycom and 'show
> hints' now shows 2 watchers.  And the Polycom doesn't show 'in use'
> state of buddy anymore.
> 
> I'd like to play with this...put together specific, reproducable
> scenarios along with debugging traces.  Once I do that, should I open an
> issue on the bug tracker?
If there's a bug, yes, please do.

/O