Skip to content

3.1 - Building Custom Cluster Machine Images

Overview

  • Tanzu Kubernetes clusters = Workload Clusters
  • Applications are typically deployed to these clusters and managed by TKG Management clusters
  • Provisioning and configuring Tanzu Kubernetes clusters is the key functionality that TKG provides

Learner Objectives

  • Describe the steps to build a custom image
  • Describe the available customizations

Deploying Clusters Using Custom Machine Images

  • Custom machine images can be used as VM templates for Tanzu Kubernetes workload cluster nodes
  • Each custom machine image packages a base OS version and a Kubernetes version, as well as any additional customizations into an image capable of running on:
  • vSphere
  • AWS
  • Azure

  • Base OS can be any OS that VMware supports but does not distribute e.g. Red Hat Enterprise Linux (RHEL) v7

  • Note: for TKG v1.3 - only workload clusters support custom images, both these and management clusters are supported in v1.4

Custom Machine Image Customizations

Untitled

  • Variables including the above can be used to customize the image builds.

Building Custom Images - Requirements

  • The following requirements must be met for custom images:
  • Account on vSphere/AWS/Azure with sufficient permissions
  • Docker installed on Mac OS or Linux workstation
  • For AWS - aws CLI enabled
  • Azure - az CLI enabled
  • vSphere - OVFTool must be installed

Steps to Create Custom Image

  1. Determine and download image builder configuration version to be built from Each version corresponds to the Kubernetes version that the Image Builder will use.
  2. For vSPhere - load OVF tool in the image builder container image
  3. Prepare the required configuration
  4. Run image builder (recommended to do a dry run first)

Using Custom Images

  • After creating the custom images, you need to allow the Tanzu CLI to use the image - this requires creation of a custom Tanzu Kubernetes release based on the image

  • Open the BOM (Bill Of Materials) file corresponding to the Kubernetes version being used by the custom image

  • Edit the configuration as appropriate
  • Encode the BOM configuration in Base64
  • Create a ConfigMap YAML to the tkr-system namespace
  • Save the ConfigMap, set the context to the management cluster the Tanzu Kubernetes release is to be added to, and apply via kubectl apply
  • Verify the custom Tanzu Kubernetes release was added via tanzu kubernetes-release get or kubectl get tkr