Task cluster scheduling management method

TL;DR: In this article, a task cluster scheduling management method is proposed, where a server node for executing task scheduling is separated into a scheduling center and an executor, and the scheduling center comprisesa scheduling pool, a Quartz scheduler, a callback service and a scheduling log, and is in charge of managing scheduling information and sending out a scheduling request according to configuration of the scheduling information; the executor comprises a server, a task execution queue and a task object.

Abstract: The invention discloses a task cluster scheduling management method. A server node for executing task scheduling is separated into a scheduling center and an executor; the scheduling center comprisesa scheduling pool, a Quartz scheduler, a callback service and a scheduling log, and is in charge of managing scheduling information and sending out a scheduling request according to configuration of the scheduling information; the executor comprises a server, a task execution queue and a task object; the server is in charge of receiving heartbeat detection information and task triggering information of the scheduling center; and the executor receives task information of the scheduling center through the server, packages the task information into the task object, adds the task object to the task execution queue for waiting for execution, starts to execute the task object when execution time is up, and asynchronously returns a final task execution result to the scheduling center through a specified interface of the scheduling center. The scheduling center is decoupled from tasks, so that system availability and stability are improved and the performance of a scheduling system is no longer limited by the tasks; and the tasks focus on the execution operation, so that development and maintenance are simpler and more efficient.