This topic lists the latest changes to resource-controller.

Introduction

resource-controller is the core component that is used to dynamically schedule pods. If you want to enable topology-aware CPU scheduling for a professional Kubernetes cluster, you must install resource-controller.

Instruction

For more information about how to use resource-controller, see Topology-aware CPU scheduling.

Release notes

April 2021

Version Image address Release date Description Impact
v1.2.0-ec8a979-aliyun
  • registry.cn-hangzhou.aliyuncs.com/acs/node-resource-agent:v21.3.9.0-adecd8a-aliyun
  • registry.cn-hangzhou.aliyuncs.com/acs/resource-controller:v1.2.0-ec8a979-aliyun
2021-04-20 resources-controller is compatible with containerd. Access to multiple containers and their metadata is supported. Container logs can be collected. The resource usage of tasks on a node can be adjusted in real time based on task priorities. No impact on workloads.

March 2021

Version Image address Release date Description Impact
v1.1.0-e7388cb
  • registry.cn-hangzhou.aliyuncs.com/acs/node-resource-agent:v21.3.9.0-adecd8a-aliyun
  • registry.cn-hangzhou.aliyuncs.com/acs/resource-controller:v1.1.0-e7388cb
2021-03-09 node-agent is released. It allows you to add labels to ConfigMaps that store information about non-uniform memory access (NUMA) nodes. No impact on workloads.
January 2021
Version Image address Release date Description Impact
v1.1.0-e7388cb
  • registry.cn-hangzhou.aliyuncs.com/acs/node-resource-agent:121ffbe
  • registry.cn-hangzhou.aliyuncs.com/acs/resource-controller:v1.1.0-e7388cb
2021-01-26 Distributed deployments and custom values of the cpuset parameter are supported. No impact on workloads.