1. set log_checkpoint_timeout=n (n seconds) in init.ora to force data write back to redo logfile in hard disk when checkpoint occurs.
2. issue alter system switch logfile command at any time to force redo logfile switch. then the closed redo logfile can be archived.