A virtual node is a scheduling capability provided by an elastic cluster. When an elastic cluster is created, a virtual node is created in each subnet where the container network is located.
You can create a virtual node in the following use cases:
When you run a large scale of Workloads in an elastic cluster, a Pod cannot be created due to the exhaustion of IP resources of the subnet in the AZ where the service is located. We recommend you create virtual nodes in a new subnet to obtain more IP resources.
As service scale expands, services need to be automatically broken down and distributed in multiple AZs. We recommend you create virtual nodes in a new subnet to expand the resource AZ.
When you create a Workload, Pod pending will be caused by resource shortage in the AZ. This may be reflected in the following cluster events:
EVENT REASON : "FailedCreatePodSandBox"
EVENT MESSAGE : "Failed to create pod sandbox in underlay (will retry): insufficient resource"
You can create virtual nodes associated with other AZs to expand the available resources in the cluster.
Virtual nodes are not billed, and fees will be charged based on the CPU, GPU, and memory applied for a workload as well as the workload running time. For more information, see Billing Overview, Product Pricing, and Purchase Limits.
This document describes how to create virtual nodes in an elastic cluster in the TKE console.
After a virtual node is created, see Managing a Virtual Node for other operations.
Was this page helpful?