Container cluster scheduling method and device, equipment and storage medium

Abstract

The invention discloses a container cluster scheduling method and device, equipment and a storage medium, and the method comprises the steps: collecting the resource utilization condition of each node in a container cluster, creating a training set based on a collected sampling data set, and constructing a corresponding node available resource prediction model based on a principal component analysis method through employing the training set; matching available resources of each node in the node available resource prediction model with the current to-be-processed container load request to determine a target node from each node; and scheduling the container load request to the target node through the scheduler, so that the target node processes the container load request. According to the method, the node of which the available resources are most similar to the current container load request is determined from the nodes by using the constructed node available resource prediction model, and the node is used as the target node to process the container load request, so that the elastic adjustment of the allocation of the container load request is realized, and the node resource utilization rate is improved.

Description

technical field [0001] The present invention relates to the field of computer technology, and in particular, to a container cluster scheduling method, device, device and storage medium. Background technique [0002] At present, the rapid development of cloud computing makes it a mainstream to choose cloud platform as the deployment platform for application deployment. In order to adapt to dynamic and changing workloads, application scenarios and service quality requirements, cloud platforms need to dynamically adjust workload resources in a scalable manner. [0003] In the prior art, the scheduler scheduling algorithm of the Kubernetes cluster in the container technology includes two parts of the scheduling algorithm predicates and priorities. The scheduler scheduling algorithm first executes the predicates algorithm to filter some nodes (nodes), then executes the priorities algorithm to score all nodes, and finally selects the best node with the highest score from all node...

AI Technical Summary

Problems solved by technology

The scheduling algorithm of the Kubernetes cluster is more calculated based on its own cr distribution and node affinity. For resource occupancy, only idle resources and resource balance are calculated. This algorithm is prone to resource fragments and cannot maximize utilization

Images