A method, system, device and storage medium for optimizing timing tasks

Abstract

The present invention provides a method, system, device and storage medium for optimizing timing tasks. The method includes: starting multiple working threads to monitor and consume timing task messages, and binding the working threads to corresponding message queue index subscripts set; pull the timing task message corresponding to the subscript of the message queue index from the task cache queue, move the timing task message to the preprocessing cache queue, and send the timing task message to the service end according to the event identifier; in response After the business end receives the scheduled task message and consumes it successfully, the business end sends a transaction submission instruction; and in response to the successful transaction submission, deletes the scheduled task message from the preprocessing cache queue. The invention reduces the load pressure of the main storage node, reduces the system processing time delay, and improves the concurrency performance.

Description

technical field [0001] The present invention relates to the field of distributed storage systems, and more specifically, to a method, system, device and storage medium for optimizing timing tasks. Background technique [0002] With the large-scale popularization and application of digitalization, the distributed storage system as the underlying support has sprung up like mushrooms, and scheduled tasks and related optimization have become an important part of the complete life cycle of storage products. Generally, the timing task processing mostly adopts the following processing flow: the timing task is persistently stored in the timing task database; the timer is used in the process body to scan the database regularly, and after finding the task to be processed, the timing task is placed in the memory queue of the node and waits for Hand it over to the specific business processing module for consumption processing, and update the task status after the business processing mod...

AI Technical Summary

Problems solved by technology

When the amount of data in a single table of the database is huge, it will take a long time even to scan all pending scheduled tasks from the data table

[0005] (3) In the case of node failure scenarios in distributed large-scale clusters, the process has a weak ability to deal with failures and cannot automatically handle and recover failures from the inside. It needs to use external scripts to start a new process to handle failure scenarios separately. The fault tolerance is poor, and failures The processing capacity is weak, and it is prone to large delays under the fault of a large amount of tasks to be processed

Images