Description
A N1QL query submitted to Spark will currently end up on a random Spark worker node.
To support Spark workers being co-located on Couchbase nodes, it would be better to find the Couchbase nodes that are running query services, and try to run the query on those Spark workers.