Distributed cluster database synchronization method and device, equipment and storage medium

Abstract

The invention discloses a distributed cluster database synchronization method and device, equipment and a storage medium. The method comprises the steps that a main management node persists database information to a local system disk and backs up the database information to a main management node BMC system in an in-band communication mode; in response to the fact that the main management node BMCsystem detects that the main management node fails or is switched, backup database information in the main management node BMC system is sent to a standby management node; and the backup management node updates a database of the backup management node by using the backup database information. According to the scheme, the database information of the main management node is backed up to the main management node BMC system, and the backup management node updates the database of the backup management node by adopting the backup database information of the main management node BMC system, so thatthe database information is prevented from being lost, the reliability and integrity of the database information of each node in a cluster are effectively ensured, manual intervention synchronizationis reduced, and maintenance cost is reduced.

Description

technical field [0001] The present invention relates to the field of distributed storage, and more particularly relates to a method, device, computer equipment and storage medium for synchronizing a distributed cluster database. Background technique [0002] A distributed storage cluster system is generally composed of multiple servers to provide unified external services. In this paper, the server is represented by a "node". The traditional distributed cluster has its own web management system, including the client (Client) and the server (Server) It runs on the management node, so the database (DB) it maintains is usually updated on the management node and synchronized to other nodes periodically (for example, in a period of 30s); There will also be one or more standby management nodes. When the main management node fails (such as sudden power failure, downtime, etc.), it needs to switch to the standby management node immediately. However, the time of power failure is usua...

AI Technical Summary

Problems solved by technology

[0002] A distributed storage cluster system is generally composed of multiple servers to provide unified external services. In this paper, the server is represented by a "node". The traditional distributed cluster has its own web management system, including the client (Client) and the server (Server) It runs on the management node, so the database (DB) it maintains is usually updated on the management node and synchronized to other nodes periodically (for example, in a period of 30s); There will also be one or more standby management nodes. When the main management node fails (such as sudden power failure, downtime, etc.), it needs to switch to the standby management node immediately. However, the time of power failure is usually random and cannot be determined. At this time It may happen that the management node database has stopped running before it is synchronized. If the latest updated database information of the main management node before the power failure has not been synchronized to the standby management node, it has failed, and these database records may be lost due to inability to synchronize in time , leading to some unknown exceptions may occur after the management node is switched, causing some management flow risks, so the existing database synchronization method of distributed storage needs to be improved urgently

Images