[asterisk-bugs] [Asterisk 0012773]: [patch] Agent doesn't propagate device state changes from underlying device.

[noreply at bugs.digium.com]

A NOTE has been added to this issue. 
====================================================================== 
http://bugs.digium.com/view.php?id=12773 
====================================================================== 
Reported By:                davidw
Assigned To:                Corydon76
====================================================================== 
Project:                    Asterisk
Issue ID:                   12773
Category:                   Channels/chan_agent
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     ready for testing
Asterisk Version:           1.4.20.1 
SVN Branch (only for SVN checkouts, not tarball releases): N/A 
SVN Revision (number only!):  
Disclaimer on File?:        N/A 
Request Review:              
====================================================================== 
Date Submitted:             06-02-2008 11:30 CDT
Last Modified:              06-26-2008 05:54 CDT
====================================================================== 
Summary:                    [patch] Agent doesn't propagate device state changes
from underlying device.
Description: 
Given a SIP phone logged in with AgentLogin, and monitored with qualify, if
you remove the power on the phone (or, presumably, unplug it from the
network) Asterisk detects that the phone has become unavailable, but the
agent continues to receive calls, resulting in callers going into a black
hole.

Ideally, I believe that the agent should detect the state change on the
underlying device and change its own state to logged off, and therefore
generate a state change to CHANUNVAIL.  This would result in correct
hinting for the agent's state.
======================================================================
Relationships       ID      Summary
----------------------------------------------------------------------
has duplicate       0012771 Bogus <member> is still 'Not in U...
====================================================================== 

---------------------------------------------------------------------- 
 davidw - 06-26-08 05:54  
---------------------------------------------------------------------- 
I got this feedback from a person able to test this:

"The disappearing agent trick seems to work much better now.

If I pull the data cable on the agents phone, wait for unreachable and
then call into the queue it just rings out. I left it for over a minute and
then plugged the data cable back in. The call was answered as soon as the
endpoint was detected as reachable.

The only wrinkle I could see was that the agent's status was still "logged
in" even though they were not reachable."

Where I have some reservations are:

- I'm not sure what the "has owner" test was trying to achieve, so I'm not
sure whether the "inherited status > -1" test is completely safe (e.g. if
it is trying to protect against accessing a partially initialised
structure, inherited status might be zero);

- I wonder if giving inherited status absolute priority is the right
thing.

However, in terms of soak testing this, currently the marketing person is
on leave and the developer is working on other aspects of the application
so isn't extensively exercising the system, so empirical evidence of any
collateral damage may be slow in coming. 

Issue History 
Date Modified   Username       Field                    Change               
====================================================================== 
06-26-08 05:54  davidw         Note Added: 0089271                          
======================================================================