(1) Short problem description:

Cisco Phone Crashes and Burns During SIP Session with multi-homed SIP endpoints

(2) Longer problem description (what happens):

Cisco phone has a tendency to crash and burn when communicating with multi-homed SIP endpoints (such as an Asterisk PBX with multiple Ethernet interfaces or IP aliases).  I do not know if Asterisk's output messages in this scenario strictly adhere to standards, but no other client I've run in the same scenario (Snom 200, X-Ten Lite, etc.) has had this problem.  Only the Cisco 7960 crashes out.  Even if the problem lies in the messages themselves, the Cisco should handle it more gracefully than a crash & reboot.

(3) Possible solution (what did you expect):

If the first IP address listed on the Asterisk PBX box is the IP address that the Cisco 7960 tries to connect to, and both are in the same subnet, all is well.  Also, if the Cisco is remotely located (away from the same Ethernet as the Asterisk PBX) and you call upon the PBX at its external IP address, there continues to be no problem.  (Assuming you use the fromdomain option in the asterisk SIP setup record for the Cisco & create a DNAT rule on the router in front of the Cisco such that the internal address destination gets rewritten as the external interface of the Asterisk server...)

(4) How to reproduce (if possible / applicable):

It should be pretty easy.  Set up a multi-homed Asterisk server by using IP aliasing and try to communicate with the server using the Cisco on the same Ethernet segment.  Configure the Cisco in an internal-only address space, but make the default gateway and the primary interface on the Asterisk server is in the internet-routable address space.  SIP signaling will work, but as soon as audio data is passed, the phone will crash.

(5) Workaround used currently (if applicable):

See above #3.

(6) The behaviour wanted exists in product (name manufacturer and product as exactly as possible):

Snom 200 ( a hardphone) , XTen Lite Soft Phone ( a softphone)

(7) Would you classify this as:

(A) Bug/malfunction: Can not deploy before fixed

Category A, in terms of a mass deployment.  I cannot imagine sending this phone out into the field with this kind of scenario outstanding. Category C, in terms of internal use in a tightly controlled network.

(9) Contact Information

Matthew Hardeman <mhast@SPAM.papersoft.com>

(10) If you have filed a DDTS/CARE case, what case number is it

There is a Cisco case EE225816

(11) Report:

Message-Id: <p052106b3bb34e3fe9dec@loligo.com>
Date: Fri, 11 Jul 2003 15:02:09 -0700