Symptom:
Calls to a Rendezvous Conferencing resources may fail intermittently under load.  The issue is seen when calls are sent in the following manner:
Endpoint -> Leaf Cluster -> SME Cluster -> Conductor

There are 2 issue here in Early offer trunk scenario:

1. Mid call Invite with new SDP received with LATM parameters is not processed properly in media layer which  causes 200 Re-INVITE to be stuck. This is addressed with Media layer fix "CSCuq01973".

2. SDP in 18x/ 200 OK  answer sent to A leg is new SDP received from Mid call offer instead of original SDP answer  when there is delay or race condition in call-control Response signal and new offer received from B leg.

Conditions:
The issue is seen when calls are sent in the following manner:
Endpoint -> Leaf Cluster -> SME Cluster -> Conductor

The clusters have to be under load where the race condition is caused by delays in the SDL Signaling queue and "m=" line in SDP contains LATM codec.