Parallel execution method and device for timed task on multiple servers

Abstract

The invention provides a parallel execution method and device for a timed task on multiple servers. The uniform configuration of server nodes can be realized, limited calculation resources are fully utilized, the atomicity of a task is supported and guaranteed, operation and maintenance cost is reduced, the problem of single point failure is solved, the stagnation of service due to the single point failure is avoided, no manual intervention is required when the server crashes, the splitting of a large task is supported, application developers only need to pay attention to specific business development, and availability and expansibility are good. The parallel execution method for the timed task on the multiple servers comprises the following steps that: receiving the attendance request ofa task living example submitted by each server node, carrying out attendance on the task living example according to a preset rule, adding the task living example which is subjected to successful attendance to a task execution list to enable the corresponding server to execute the task living example; and polling the execution state of each task living example in the task execution list, signing out the task living example which finishes being executed, and removing the task living example from the task execution list.

Description

Technical field [0001] The present invention relates to the technical field of computers and their software, in particular to a method and device for parallel execution of timing tasks on multiple servers. Background technique [0002] There are timing tasks in a large number of application systems. For example, in the actual development process of application systems, there are a large number of system background tasks that do not require manual intervention and need to be executed regularly. These timing tasks are often unavoidable to take into account single point problems, Coordination problems and split problems require a complete solution to support the effective operation of the system. Using Spring Quartz can efficiently support this timing task scheduling scenario, so that these background tasks can be automatically run at a specified time and according to a specified strategy. [0003] Quartz is an excellent enterprise-level task scheduling engine. Spring has encapsulate...

AI Technical Summary

Problems solved by technology

[0010] First, the configuration files of different online server nodes are inconsistent, which adds a lot of cost and burden to operation and maintenance. Every time online deployment, it is necessary to carefully verify the business configuration of different servers one by one; when a server fails, The work of fault location and business recovery brings additional time costs; the resource utilization among the servers is not balanced enough, and there is a problem that some servers are under high operating pressure while others are under little operating pressure

[0011] Second, there is a single point of failure problem. Once the server fails, it can only be known through the corresponding business alarm, and then the corresponding business can be restored through manual intervention.

[0012] Third, when a single server uses a single thread or a traditional thread pool for multi-threaded business processing, when a specific business performs time-consuming operations such as I / O, it often causes the CPU to idle, making the application program's computing resources Utilization rate is not high

Images