<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Aug 4, 2016 at 11:12 AM, Ryan Rittgarn <span dir="ltr"><<a href="mailto:rrittgarn@techpro.com" target="_blank">rrittgarn@techpro.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Everyone, long time reader, first time poster in this particular thread. I'm one of the holdouts still using chan_sip for day to day uses as I have too many applications and implementations that rely on it. That being said I think I have found a bug, but am not sure the best way to go about isolating it to specific code changes, but I'd love to help any way I can.<br>
<br>
We recently (today) just upgraded a test box from 13.4 to 13.10. We process inbound calls through app_queue and then use one of our apps that issues a redirect via AMI to direct the channel to a parking extension, or more specifically an extension in dialplan that calls the Park() application. Up until now our behavior was such that when an agent of the queue parked the call, the agent's channel would be destroyed upon parking (as expected) and the caller's channel would show bridged with app_park in core show channels. Now as part of our application also keeps tabs on who is on the phone via AMI events, we noticed that when a call comes in through the queue, the original agent's channel stays up until caller is done with the call and hangs up. We use the queues for our triage people, and then pass through to support staff via picking up parked calls, so some of these calls can be lengthy. Upon further investigation doing core show channels I show that the original agent's channel still shows up, but I cannot issue the channel request hangup command without error. It tab completes and everything to the full channel, but when I hit enter, I get a 'not found' message.<br></blockquote><div><br></div><div>That agent channel hanging around sounds like a channel ref leak or something still has<br>a reference to the channel that they won't let go until the call the agent channel was<br>involved with hangs up. (You actually confirm this in your description.) I expect that the<br>lingering channel has already been hung up since you got a channel not found when you<br>tried to hang it up. The channel is still visible in the stasis channel cache because a<br>reference to the channel still exists but the channel is actually hung up.<br><br></div><div>I suspect that it is app_queue that has the lingering channel reference for its queue_log.<br></div><div><br></div><div>Richard<br></div></div><br></div></div>