I've an MySQL 5.1 slave for our BI team.
They need to make some CREATE SELECT with big select queries (several million lines).
As CREATE SELECT is a DDL, if the replication attempts to update some rows in same tables than the SELECT statement, replication is blocked until the freeing of the CREATE SELECT.
Do you now a good non-blocking alternative to thoses CREATE SELECT statements?
I thought to an SELECT INTO OUTPUT FILE then LOAD DATA INFILE but they will fill out our disks as BI guys like to do... :)