Overview

Nodes are the fundamental building blocks of a cluster. Nodes can be virtual machines or physical servers, and each node runs the components required to host Pods, including kubelet, kube-proxy, and the container runtime.

Users with platform management permissions can add, manage, and plan nodes under clusters.

NOTE

Adding nodes to imported clusters or deleting nodes from imported clusters is not supported.

Node types

  • Control plane nodes run cluster components such as kube-apiserver, kube-scheduler, kube-controller-manager, etcd, container networking, and some platform management components.
  • Worker nodes host business Pods running on the cluster. Plan worker node count based on workload demand.

Node operation paths

The following tasks apply to clusters whose nodes run Alauda OS and to clusters whose nodes run a non-Alauda OS.

NeedContinue with
Plan infra nodes or custom role nodes.Infra Nodes
Tune CPU, memory, or topology manager policies for performance-sensitive workloads.Resource manager policies

Alauda OS Nodes

Nodes on Alauda OS clusters are managed through Cluster API resources and are replaced rather than changed in place.

NeedContinue with
Configure machine health checks, or scale and maintain nodes on a specific infrastructure provider.Alauda OS Nodes

Non-Alauda OS Nodes

NeedContinue with
Add nodes to or delete nodes from a cluster.Add and Remove Nodes
Manage labels, taints, and node scheduling state.Manage Nodes
Plan in-place operating system upgrades for eligible nodes.Plan an In-Place OS Upgrade for Nodes

Linux node availability check

Before building a workload cluster on a non-Alauda OS, complete the node checks and confirm that every node meets the requirements. Cluster deployment can fail when prerequisites are not satisfied.

Supported operating systems and CPU models

For supported operating systems and CPU models, see Supported Operating Systems and CPU Models.