Resource allocation method, device and server

Abstract

The invention provides a resource distribution method and device and a server, and the method comprises the steps: determining the total amount of resources in a container management cluster; Determining the ratio of the distributed resource quantity to the total resource quantity of each application in the container management cluster; Determining the application with the minimum ratio; Allocating a preset number of Pod resources to the application with the minimum ratio, and updating the allocated resource quantity of the application with the minimum ratio; Judging whether the sum of the allocated resource quantities of all the applications is greater than the total resource quantity or not; If the judgment result is no, the resources are continuously distributed through the steps; And if the judgment result is yes, allocating a corresponding amount of resources to each application according to the allocated resource amount of each application, the allocated resource amount of the application with the minimum ratio being obtained by subtracting the preset number of Pod resources from the updated allocated resource amount of the application with the minimum ratio. According to theresource allocation method and device, resources can be fairly allocated to multiple applications in the cluster, and meanwhile it is guaranteed that the allocated resources can be effectively and reasonably utilized.

Description

technical field [0001] The present disclosure relates to the technical field of virtualization, and in particular to a resource allocation method, device and server. Background technique [0002] Containers are a lightweight, portable, self-contained software packaging technology that enables applications to run the same way almost anywhere. Different from traditional virtualization technologies, containers run in a certain user space of the operating system, are isolated from other processes of the operating system, and are much smaller in size than virtual machines. Starting a container doesn't require booting the entire operating system, so containers can be deployed and started faster, with less overhead, and are easier to migrate. [0003] Kubernetes is an open source platform for automated container operations. Through this platform, containers can be deployed and replicated automatically, container scale can be expanded or contracted, containers can be organized into...

AI Technical Summary

Problems solved by technology

If there are multiple applications running in the container management cluster, it is difficult to reasonably allocate resources for each application as a whole by using the resource limit method set based on historical data or engineer experience, which may easily lead to low resource utilization of some applications, while Some applications cannot run properly due to insufficient resources

Images