Method for realizing high availability of PostgreSQL

Abstract

The invention discloses a method for realizing high availability of PostgreSQL, the method is applied to a system for realizing high availability of PostgreSQL, and the system comprises a client, an agent unit, a sentry unit, a PostgreSQL cluster and a storage unit. The agent unit, the sentry unit and the PostgreSQL cluster register own unit information into the storage unit, the agent unit is anaccess point of the client, and the agent unit is always connected to the main library of the management unit by reading the information on the storage unit; the sentry unit monitors the survival states of the master library and the slave library in real time by reading the information on the storage unit, if the master library breaks down, automatic switching of the master library fault is achieved, difference data are compensated in the switching process, and therefore it is guaranteed that data are not lost when the master library breaks down. Wherein the database cluster is a hierarchicaldatabase master-slave replication cluster. According to the invention, automatic switching of main library faults can be realized, and data loss is avoided.

Description

technical field [0001] The invention relates to the technical field of databases, in particular to a method for realizing high availability of PostgreSQL. Background technique [0002] The hierarchical database is a high-performance database based on the LSM algorithm, and the data is stored on the hard disk according to different levels. The database supports master-slave replication architecture, which is divided into master database nodes and slave database nodes. The master database generates incremental logs in real time, and the slave database synchronizes the master database incremental logs in real time. The current position of the incremental logs is called the synchronization offset; The data on the master database that is not synchronized to the slave database is called differential data, which can be obtained through the interface of the hierarchical database itself; in general, the synchronization offsets of the master database and the slave database are the sam...

AI Technical Summary

Problems solved by technology

[0009] 2) Manual processing is complex

[0010] Once the main database is down, manual processing is required, including selecting a new main database, compensating for difference data, restoring the new cluster, and modifying the client configuration. The operation steps are complex and time-consuming

[0011] 3) Data loss

[0016] 1) The main database failure cannot be automatically switched

[0017] Once the main library fails, manual intervention is required, and automatic failover cannot be achieved, and the availability is relatively low

[0018] 2) Complicated operation

[0021] The virtual IP service itself has high availability problems, which increases maintenance costs. If the virtual IP service is unstable, it will also affect the availability of the database

[0022] 4) Data loss

[0023] When the main library is down, the data that has not been synchronized to the slave library will be lost

Images