we want to build the following ambari clusters
1) one cluster include ambari cluster with all services except kafka and zookeeper ( as HDFS , YARN , namenode , resource manager , mapreduce , etc )
this cluster will be HDP version 2.6.5
2) the second ambari cluster will build from HDP 3.1 and include kafka and zookeeper machines only
both cluster should connected and works together
is it possible ?