Split brain preventing method for highly available cluster system

Abstract

The invention provides two methods for preventing a brain from splitting of a cluster system. The first method is as follows: whether the brain is split is judged by detecting whether last two heartbeat cables are cut at the same time. The second method is as follows: whether the service is started at a node of the other side is judged by a ping floating IP method, thereby preventing the brain from splitting. The two methods can effectively avoid the brain splitting problem of the dual-computer high-useable cluster system, ensure that data is not damaged, and have the advantage that the support of any other hardware is not needed.

Description

technical field

[0001] The invention belongs to the technical field of computer clusters, and in particular relates to two methods for preventing the split-brain problem of cluster systems, so as to ensure the high reliability of the system. Background technique

[0002] With the rapid development of communication network technology, key areas such as telecommunications, finance, and e-government have higher and higher requirements for server availability. Stopping service due to server failure will cause huge losses. Using a high-availability cluster system can greatly improve The availability of the system minimizes the impact of failures caused by software, hardware, and human beings on the business.

[0003] The dual-machine high-availability cluster system consists of two servers called nodes connected through the network (or other connection methods such as serial ports). Among these two servers A and B, A is the working machine, and B is the backup machine, and they ...

Images