Github user mattyb149 commented on a diff in the pull request:
https://github.com/apache/nifi/pull/3167#discussion_r232697514
--- Diff:
nifi-nar-bundles/nifi-standard-bundle/nifi-standard-processors/src/main/java/org/apache/nifi/processors/standard/GenerateTableFetch.java
---
@@ -114,6 +114,7 @@
+ "max value for max value columns. Properties should be added in
the format `initial.maxvalue.<max_value_column>`. This value is only used the
first time "
+ "the table is accessed (when a Maximum Value Column is
specified). In the case of incoming connections, the value is only used the
first time for each table "
+ "specified in the flow files.")
+@PrimaryNodeOnly
--- End diff --
Since GenerateTableFetch can accept incoming flow files, I don't think we
should restrict this to run only on the primary node. For example, you could do
a ListDatabaseTables, then distribute the flow files among the cluster, where
each node's downstream GTF could do the fetch in parallel. In fact that's the
main reason we have GTF rather than just QueryDatabaseTable.
---
