Hi Tony,
If UCCX Publisher (primary Engine) fails and Subscriber (second Node) takes over the Engine mastership (assuming that the CUCM is up and running during this time and no failover is happening in CUCM cluster) than,
1. All the connected calls will continue to work, eventhough you will see a warining message in the CAD and CSD stating that the active conection to th UCCX first node is lost, the RTP is kept alive but you may not see the exact call detail recordes because Primary engine fails here. However the CAD and CSD will automatically get connected to the UCCX second Node.
2. New calls attempetd during this failover time will not get connected, you need to wait untill the secondary nodes takes control fully.
3. You can not perform any config changes if any one of the UCCX node is down in a High Availability setup.
4. IPPA Agents need to login one more time with the UCCX second node service selected.
Hope this helps.
Anand
Please rate helpful posts !!