Spark Cluster pods spawning with affinity/antyaffinity/nodeSelector/nodeName

Currently it is not possible to set affinity/antyaffinity/nodeSelector/nodeName on master and workers pods in cluster.
It would be nice if the spark-operator had a feature to choose Kubernetes nodes where to schedule spark cluster.
