When attempting to use the REST API to start new instances within Process Automation it will work immediately after starting the Orchestrator cluster nodes, but if they are restarted REST Calls return the error "Feature 'MdrElementID' not found"
We are seeing that when the the cluster nodes attempt to change the Master Role, which happens whenever the current Master node is shut down, the Orchestrators seem to be properly passing the Master role and is attempting to start the REST API on the new Master node, but the Catalyst engine which is serving the REST API does not properly start the API which results in REST API Calls to return the error "Feature 'MdrElementID' not found" which is returned to the REST call.
Unfortunately, CA Technologies has ceased development on the Catalyst engine some time ago and the personnel responsible for this offering are no longer with CA Technologies. At this point as Catalyst is no longer a viable option for serving our REST API we will be replacing it in the next release of Process Automation.
As a workaround until 4.3 sp2 is available in a couple of months; anytime any Orchestrator node is stopped, and especially anytime the ‘Master’ node is stopped, ALL orchestrator nodes should be stopped, to allow the first to start, assume the master role and correctly service the REST API