The Cisco IP Phone 7940 in a remote office rings continuously without any successful call. The remote office is connected to the central office through the VPN connection (PIX VPN). This problem occurs after upgrading to Cisco CallManager 3.3(3). Calls sourced from the remote office to the central office are fine; both ends can talk with the Real-Time Transport Protocol (RTP) stream being exchanged well and then hang-up normally.
However, when the central office Cisco IP Phone calls the IP Phone at the remote office, the call does not get established. Although the handset or speaker phone is picked up, it rings constantly. Unplugging or picking up and hanging up the Cisco IP Phone for a couple of times stops the ringing. These symptoms are observed only on Cisco IP Phone in one location. Phones in others locations work well. Initially, it was thought that the problem was with 7940s since this location had a 7940 and other locations had 7960s. However, when a 7960 was sent to the remote location, it exhibited the same problem.
The Cisco CallManager trace shows that the call is established and the media transmission is fine. It also thinks that RTP was established between the two Cisco IP Phones and nothing unusual is detected in the trace.
There are several possible causes for continuous ringtone, usually associated with a disconnect problem with FXO gateway ports. For a more general description on this problem refer to Understanding FXO Disconnect Problem.
Clearly, the symptoms of this particular problem are not associated with an FXO problem since it is end to end IP. In addition, the symptoms are unusual since in this case an upgrade to CallManager caused a problem in an IP phone in only one remote location. This fact, leads one to investigate differences in this particular remote location. In issues of voice signaling, a usual suspect is firewalls.
In this case, the issue was determined to be with the PIX Firewall. The remote office is running PIX 6.2(2) whereas the central site has 6.3(1). Uprading the PIX OS on the remote office to 6.3(1) solved the issue. Although it was not fully determined what caused the problem, we can at least raise an awareness here that with the combination of 6.2(2) and CallManager 3.3(3), there may be a problem with continuous ringback.