If I use the CONNECT engine to federate data from another table, and then issue something like the following on the CONNECT table:

ALTER TABLE tbl1 ENGINE=InnoDB

When converting from CONNECT to InnoDB, both the schema and table contents will be copied.

If I instead use the SPIDER engine to federate the access, only the schema definition is modified with the previous SQL statement, but no data is copied into the local table, instead the table remain empty.

My particular intent is to archive legacy data, moving it from one physical server to another after it has aged out.

I would simply use the CONNECT engine itself, but it fails to support SETs, ENUMs, and NULLable indexes.