Method, terminal device and storage medium for data disaster recovery playback of computing node

Abstract

The invention discloses a method for data disaster tolerance playback of computing nodes, including the following processes: a process of setting a disk directory, a process of setting the frequency of checkpoint file backup, a process of data entering a computing node, a process of data being processed by the computing node, and a process of computing The process of data playback when the node is abnormal. A terminal device for disaster recovery and playback of computing node data, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements a computing node when the processor executes the computer program The steps of the method for data disaster recovery playback. A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements steps of a method for data disaster recovery playback of a computing node. It not only realizes the high-performance data disaster recovery of computing nodes, but also solves the problem of timeliness in data recovery and playback.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a method, terminal device and storage medium for data disaster recovery and playback of computing nodes. Background technique [0002] With the advent of the era of massive data, data computing power and data disaster tolerance have become indispensable indicators in software system architecture. As for the computing power of data, the horizontal expansion of computing nodes can be achieved by means of distributed load balancing at the system architecture level, so as to achieve an infinite improvement of computing power. However, how to ensure the disaster recovery and playback of data on each computing node has also become the focus of the industry. [0003] Most of the data disaster recovery on traditional computing nodes is to use WAL (write-ahead-log) for the data cached in the queue, that is, to record a log to a disk file before inserting the data into the queue. ...

AI Technical Summary

Problems solved by technology

Although this method ensures the high availability of computing node data, when writing records to disk files, because they are written to a disk file uniformly, there will inevitably be I / O performance problems of concurrent writing. In addition, when In an abnormal situation, when it is necessary to restore the data of the computing node, it is necessary to perform "full playback", that is, to load all the records in the entire disk file one by one. When the computing node runs for a long time and the calculated data is large , the history records of the entire disk file are very large, and most of the records may be useless outdated records. Performing "full playback" not only does a lot of useless work, but also the efficiency of playback is extremely low

Although all data can be recovered, the recovery time is often not satisfactory

Images