On Wed, 2009-10-21 at 15:43 -0400, Mark Gertsvolf wrote: > I had a very funny and equally annoying experience today. I have a > US-based DID number purchased from voip.ms. This number is forwarded to > an extension of sipXecs system and from there a call forwarding rule via > voip.ms sends the call to my cell phone. I received a call to the DID > this morning and answered it on my cell phone. The caller had the wrong > number so the call was over in a few seconds. After that I started > getting strange calls from the same caller every half an hour. When > answering the call I would hear dead silence on the line. This continued > for hours. Loud noise from banging the phone on the desk and use of > strong language would not stop the guy from calling over and over. > > It turns out during call disconnect of the first call sipXbridge messed > up the call termination, probably due to > http://track.sipfoundry.org/browse/XX-6698. It was then left thinking > that it had a call in progress and session timers audit would kick in > every 30 minutes. Unfortunately the ITSP SIP UA (User-Agent: > VoIPMS/SERAST) is not behaving properly and is treating the in-dialog > session timers reINVITE as a new call. The ITSP is actually making a > call to my cell phone each time. It is also OK with reusing the old "to" > and "from" tags and callId for the new call. The other piece of the > puzzle is that every time I would answer the call on my cell phone and > hang up the BYE is generated by the ITSP, but sipXbridge would reject it > with "500 out of order" response, and sipXbridge would keep the call > alive.
Wow. > Clearly treatment of the reINVITE as a new call is an issue with the > ITSP. But, I may have a hard time getting the ITSP to fix this problem > on their end. > > Is there anything we can do at the sipXbridge end to help minimize the > damage in this scenario? > > I see two opportunities: > 1. During re-INVITE processing ITSP sends provisional 180 ringing > response to sipXbridge. Could this be the hint that the UA is setting up > a new call? This seems like a possibility - an in-dialog reinvite shouldn't ever get any 1xx response other than 100. > 2. When ITSP sends a BYE to sipXbridge could this be a strong enough > indication that this call is better be disconnected, even though > sipXbridge sees that the sequence number is wrong. Allowing an out-of-sequence BYE to terminate a call doesn't seem like a good idea... _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev sipXecs IP PBX -- http://www.sipfoundry.org/
