Distributed task scheduling method and system

Abstract

The application discloses a distributed task scheduling method and system. A task trigger puts a target task, which meets a trigger condition, into a task queue, and the target task is a time-driven-class task and/or an event-driven-class task; a task scheduler obtains task scheduling information of the target task after receiving notification information issued by the task queue, loads resource requirement configuration information corresponding to a target task type, and initiates a resource requirement application, which carries the resource requirement configuration information, to a resource scheduler; the resource scheduler initiates a resource unit creating application, which carries the resource requirement configuration information, to a resource management platform according to the resource requirement application; and the resource management platform creates a task running unit of the target task according to the resource requirement configuration information in the resourceunit creating application; and a task executor controls the task running unit to locally run the target task. Unified scheduling of different types of tasks is realized.

Description

technical field [0001] The present invention relates to the technical field of task scheduling, and more specifically, to a distributed task scheduling method and system. Background technique [0002] The resource level of the traditional business support system of the entire operator is supported by minicomputers. The background batch processing tasks of the entire operation support system generate various task application frameworks according to time and events, such as time-based tasks and periodic task processing framework Task. The framework TF for real-time processing tasks, the event-driven foreground file import task processing framework BP, the workflow task event-driven scheduling framework VM, etc. [0003] The types of task framework scheduling and application running frameworks are different. Currently based on the cluster deployment of minicomputers, the number of machines in the entire cluster is large due to the large processing capacity of the machines, and...

AI Technical Summary

Problems solved by technology

Currently based on the cluster deployment of minicomputers, the number of machines in the entire cluster is large due to the large processing capacity of the machines, and the scale of the entire cluster is within the controllable range. Various tasks are individually deployed on each machine with corresponding scheduling frameworks and specific execution task processes. There are thousands or even tens of thousands of processing tasks. The deployment, operation and maintenance of these tasks on the minicomputer are all handled manually, and the challenges of operation and maintenance are becoming increasingly severe.

Images