Hypershift Toolkit

Overview

The hypershift toolkit is a set of tools and files that enables running OpenShift 4.x in a hyperscale manner with many control planes hosted on a central management cluster. This tool was jointly developed by RedHat and IBM.

Getting Started

Install on standalone environment

• Run make build to build the binary
• Construct a "cluster.yaml" to define custom parameters for the cluster. Example found here: cluster.yaml.example
• Construct a "pull-secret.txt" to provide authentication to pull from desired docker registries. Example found here: pull-secret.txt.example
• Construct and run the render command, with optional fields below: ./bin/hypershift render
  • output-dir: Specify the directory where manifest files should be output (default ./manifests)
  • config: Specify the config file for this cluster (default ./cluster.yaml)
  • pull-secret: Specify the pull secret used to pull from desired docker registries (default ./pull-secret.txt)
  • pki-dir: Specify the directory where the input PKI files have been placed (default ./pki)
  • include-secrets: If true, PKI secrets will be included in rendered manifests (default false)
  • include-etcd: If true, Etcd manifests will be included in rendered manifests (default false)
  • include-autoapprover: If true, includes a simple autoapprover pod in manifests (default false)
  • include-vpn: If true, includes a VPN server, sidecar and client (default false)
  • include-registry: If true, includes a default registry config to deploy into the user cluster (default false)
• Apply all the generated resources to the cluster kubectl apply -f output-dir/

Installing on AWS

• Install an Openshift 4.x cluster on AWS using the traditional installer
• Run make hypershift-aws on this repository
• Setup your KUBECONFIG to point to the admin kubeconfig of your current AWS cluster (ie. export KUBECONFIG=${INSTALL_DIR}/auth/kubeconfig)
• Run ./bin/hypershift-aws install NAME to install a new Hypershift cluster on your existing AWS cluster. The NAME parameter will be used to create a namespace on your existing cluster and place all control plane components in it. Infrastructure will be created on AWS to support your new cluster instance, including:
  • Network Load Balancers for API, Router, VPN
  • DNS entries for API, Router, VPN
  • Worker machine instances for your new cluster

Uninstalling on AWS

• Setup your KUBECONFIG to point to the management cluster
• Run ./bin/hypershift-aws uninstall NAME where NAME is the name you gave your cluster when installing.