Cluster expansion method and apparatus thereof

Abstract

An embodiment of the invention discloses a cluster expansion method and an apparatus thereof. The method comprises the following steps of acquiring a target number of main nodes after expansion; according to the target number and a preset proportion, determining a maximum number of the main nodes contained by a same physical machine after expansion; determining the current physical machine containing the number of the main nodes to be a target physical machine, wherein the number of main nodes is less than the maximum number; based on a first number of the main nodes contained by a first physical machine and the maximum number, determining a second number of the main nodes which need to be distributed currently; distributing the second number of main nodes in the target physical machine; determining whether to complete cluster expansion; and if the cluster expansion is not completed, returning to an execution step and determining the current physical machine to be the target physical machine, wherein the number of the main nodes contained by the current physical machine is less than the maximum number. By using the scheme in the invention, after transverse expansion of a cluster isperformed, the whole cluster can still work normally when any one physical machine in the cluster generates a fault so that all the main nodes contained by the physical machine go down.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a cluster expansion method and device. Background technique [0002] Cluster expansion methods include vertical expansion and horizontal expansion. Vertical expansion refers to increasing the capacity of some existing master nodes in the cluster. Horizontal expansion refers to adding new master nodes to increase the capacity of the entire cluster. In the prior art, when a cloud computing management platform such as openstack (an open source cloud computing management platform project) management platform is used to horizontally expand the cluster, it will determine which host node the new master node is assigned to based on the storage load on each physical machine. on the physical machine. [0003] For clusters without central nodes, such as redis (an open source Key-Value database) cluster, in order to ensure high availability of data, such clusters set up a master-slave mod...

AI Technical Summary

Problems solved by technology

[0004] However, when the cluster expands horizontally, the number of master nodes contained in a certain physical machine is too large, so that the ratio of the number of master nodes in the cluster to the total number of master nodes in the cluster does not exceed For the above preset ratio, if a failure of the physical machine causes all the master nodes on it to go down, the ratio of the number of normally working master nodes in the current cluster to the total number of master nodes in the current cluster cannot reach the above preset ratio after the failure occurs Ratio, so that the slave node corresponding to the downtime master node can no longer be upgraded to the master node, which leads to the unavailability of the entire cluster

Images