My question regarding which query engine will be using when connecting to let’s say MySQL source. Will be query on MySQL leveraging Dremio’s parallel executor nodes, or will my query will be taking resources from MySQL source?
It’ll use both. The planner will re-write all parts of the query that can pushdown into MySQL into a single SQL statement executed in MySQL. Other parts of the query that say join data from other sources will be done in parallel using Dremio’s executor nodes. The SQL that is pushed down into MySQL will be execute and fetched in a single thread. All this information is easily seen in the query profiles.