We have a schema with 10tables and currently all 4 users have ALL access.  When I (or any other user) spin up a serverless SQL warehouse, I am able to query one of the tables (million rows) in SQL Editor and get a response within seconds.  `select count(*) from schema.tbl_name`.  When I create a compute cluster with 2nodes, and run the same query in a notebook with the attached compute custer using spark.sql("..."), it just spins and spins and spins and never get a result.  Creating a similar cluster with just the driver and not the workers (Single node cluster)  works!

Serverless SQL warehouse & Single-node cluster - works in seconds

Multi-node cluster - spins with no result in 60mins