Resource allocation method, device and electronic equipment in a mesos management cluster

Abstract

The embodiment of the invention provides a resource allocation method and device in a Mesos management cluster and an electronic device. The method comprises the steps of determining at least one task which is currently in a performance state; determining the optimum resource quantity required for performing the at least one task according to the practical resource quantity used for performing the at least one task within a preset time period; calculating the total target resource quantity required for performing the at least one task according to the determined optimum resource quantity; and allocating resources according to the target resource quantity, thereby performing the at least one task. According to the embodiment of the invention, the practical resource quantity is smaller than appointed resource quantity and the obtained optimum resource quantity is also smaller than the appointed resource quantity, so the target resource quantity is smaller than the sum of the appointed resource quantity corresponding to each task; and the resources are allocated according to the target resource quantity, thereby performing the at least one task, so the resources are saved, the allocation of the resources satisfies the resource demands of the tasks well, and the resource utilization rate in the Mesos management cluster can be improved.

Description

technical field [0001] The present invention relates to the field of electronic technology, in particular to a resource allocation method, device and electronic equipment in a Mesos management cluster. Background technique [0002] The Mesos (distributed system kernel) management cluster is composed of multiple servers and is used to perform tasks for external terminals. [0003] Specifically, the Mesos management cluster includes a master node server ("master node" for short) and a slave node server ("slave node" for short), and each master node is connected to multiple slave nodes. Each slave node itself has certain resources for performing tasks (such as CPU, memory, etc.), and reports the current resource usage to the master node in real time. The master node can receive the task execution request sent by the terminal, and schedule its slave nodes to execute tasks according to the resource usage of its slave nodes. [0004] In the prior art, when the terminal sends a t...

AI Technical Summary

Problems solved by technology

[0005] In the process of implementing the present invention, the inventor found that at least the following problems exist in the prior art: the tasks performed by the slave nodes are usually long tasks, for example, it may take a month for the slave nodes to perform a certain task

In actual situations, the amount of resources specified by the terminal for a certain task is usually greater than the actual amount of resources used by the slave node to execute the task, and when the slave node executes the task for a long time, the actual amount of resources used varies with time. Dynamically changing, the actual amount of resources will be in a low usage state for a long time. Therefore, when the actual amount of resources is low, the resources allocated from the slave nodes will be wasted a lot, thereby reducing the resources in the Mesos management cluster Utilization

Images