Kubernetes资源调度策略  

Kubernetes resource scheduling strategy

作　　者：高荣 谢晓兰[1] 刘亚荣[1] 劳榕 GAO Rong;XIE Xiao-lan;LIU Ya-rong;LAO Rong(College of Information Science and Engineering,Guilin University of Technology,Guilin 541006,China)

机构地区：[1]桂林理工大学信息科学与工程学院,广西桂林541006

出　　处：《计算机工程与设计》2024年第6期1896-1902,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(62262011);广西自然科学基金项目(2021JJA170130)。

摘　　要：在Kubernetes中可以通过nodeSelector在不同节点上部署不同的应用,但是当指定节点资源不足时,应用将会处于无法被调度的状态。针对这个问题,提出一种Kubernetes资源调度策略,设计实现一个自定义调度器,在指定节点部署应用资源不足时,将等待调度的应用调度到Kubernetes集群中的其它节点上。实验结果表明,该自定义调度器的调度策略能够在指定部署节点资源不足时将等待调度的应用调度到其它节点上,提高了集群资源利用率和集群负载均衡,降低了用户应用等待调度时间。In Kubernetes,different applications can be deployed on different nodes through nodeSelector,but when the specified node resources are insufficient,the application will be in a state that cannot be scheduled.To solve this problem,a Kubernetes resource scheduling strategy was proposed,and a custom scheduler was designed and implemented.When the applications resources deployed by specified node were insufficient,the applications waiting for scheduling was scheduled to other nodes in the Kubernetes cluster.Experimental results show that the scheduling strategy of the customized scheduler can dispatch the pending applications to other nodes when the node resources specified deployed are insufficient,which improves the cluster resource utilization and cluster load balancing,and reduces the waiting time of applications.

关 键 词：云计算 Kubernetes集群 容器 资源调度 调度器 集群负载 节点选择器 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]