Design method for improving cluster business availability by using cluster management node as two computers

Abstract

The invention provides a design method for improving cluster business availability by using cluster management node as two computers. In the method, a management node two-computer manner is utilized to realize uninterrupted monitoring and management on the cluster business and shield service interruption caused by single management node shutdown, so that the cluster operation management has disaster holding capability and realizes high availability. According to the design method by using management nodes as host and slave computers, data is allocated to the mirror image disks of the two computers, so that real-time synchronization of the middle business data on the host computer and slave computer is realized; the host computer provides cluster monitoring management and job submission interface, and when the host computer is shut down, the slave computer takes over the cluster business progresses in an uninterrupted manner, so that jointless migration of the exterior service is realized; and an uniform exterior service port is provided between the host computer and the slave computer.

Description

technical field [0001] The invention relates to the technical field of computer back server, in particular to a design method for improving the usability of cluster services by using cluster management nodes as dual machines. Background technique [0002] Server dual-machine is a way to achieve business continuity. The master / standby deploys the same application service software, shares intermediate business data and external service ports, sets the master / standby to detect the heartbeat connection of each other, and the system sets corresponding migration strategies according to the resource service level. When the main server needs to be shut down for maintenance or unexpected downtime, the standby machine can take over the business of the main machine without interruption, continue to implement internal management and scheduling, and provide external service output. After the main machine recovers, the business of the standby machine can be migrated back to the main mach...

AI Technical Summary

Problems solved by technology

With the increase of cluster node business volume and the need to run more monitoring and management software, the load of management nodes is no less than that of general computing nodes. Long-term high-load work and the particularity of management tasks require us to pay great attention to the management nodes themselves The running status of the cluster, once there is a downtime, the entire cluster will lose supervision, and the cluster business will be affected

Images