An administrator is reviewing the /tmp/hacmp.out after an unsuccessful fallover and notice that  the event stop_server PROD_DB failed.  Besides manually intervening，what step must be performed to return the cluster to a stable state？（）