Abstract: To the flaws of managing scheduled task, a kind of high-availability managing scheduled task method is proposed. It uses cluster technology, which don’t lose scheduled tasks and recover running as soon as possible when hardware or software breaks down. On the basis of researching the crontab and at in Linux, it improves the accuracy of tasks execution, implements the unity of managing periodic and aperiodic scheduled task by classifying scheduled tasks according to execution time. The prototype experiments show that the method has higher availability and performance.

Key words: scheduled task, periodic, aperiodic, high-availability, cluster