Use an existing data store during migration

If an existing Jackrabbit Classic deployment uses a data store, it should be possible to (optionally) retain it also over the migration to Oak. This option should give a massive boost to migration speed especially for big repositories, as there will be no need to copy over the contents of binary values.