@Pierre Villard thank you for your continued engagement in this conversation. 1) It sounds like by default ExecuteSQL does not provide coordination between nodes for a query, to ensure the query is not duplicated on nodes; and I, as a developer, have to ensure that no duplication occurs (but I don't know what it is that I need to do to make ExecuteSQL work correctly in a cluster); so, an alternative is the combination of GenerateTableFetch and QueryDatabaseTable processors, which work properly by making the GenerateTableFetch run as an "Isolated Processor" on the primary node and QueryDatabaseTable on all nodes; Is that a correct summary of what we discussed thus far ? 2) Also you noted in your last comment : "ExecuteSQL can achieve the same kind of things, it really depends of your use case and how you define your queries." My use case is to query Oracle and SQL Server databases, both of which support parallelization of queries; how do I need to define my queries differently for ExecuteSQL to work in a cluster while querying a relational database that supports query parallelization. 3) Is there any documentation on which NiFi processors work differently in a cluster versus a single node environment and what modifications need to be made for the processors to work correctly in a Cluster. So far my understanding is ListenX , GetX, and ExecuteSQL processors work differently in a Cluster. Again, thank you for your time and patience.
... View more
