OpenShift 4.0 libvirt on GCP

Create an OpenShift 4.0 cluster in a single GCP instance.

Create a cluster

If the images are built and available in your project/zone, all you need to get started is the gcloud CLI tool.

First, launch an instance.

$ gcloud compute instances create $INSTANCE \
  --image-family openshift4-libvirt \
  --zone us-east1-c \
  --min-cpu-platform "Intel Haswell" \
  --machine-type n1-standard-8 \
  --boot-disk-type pd-ssd --boot-disk-size 256GB \
  --metadata-from-file openshift-pull-secret=openshift-pull-secret.json

Connect to the instance using SSH and create a cluster named nested.

$ create-cluster nested

Interact with your cluster with oc.

Updating the installer

Tools can be updated right from the instance itself.

Update the OpenShift installer from https://github.com/openshift/installer.git master using:

$ update-installer

Update the OpenShift installer from https://github.com/repo-owner/installer.git branch using:

$ update-installer repo-owner branch

Update the RHCOS image using:

$ update-rhcos-image

Images

Images are built with Packer. Override variables as necessary.

Source image

The source image is centos-7 with nested virtualization enabled.

$ packer build openshift4-libvirt-source.json

To override any default variable value, for example, Google Project ID:

$ packer build -var 'project=your-google-project-id' openshift4-libvirt-source.json

Provisioned image

The provisioned image implements all the OpenShift libvirt HOWTO requirements.

$ packer build openshift4-libvirt.json

To override any default variable value, for example, Google Project ID:

$ packer build -var 'project=your-google-project-id' openshift4-libvirt.json

Advanced usage

It's possible to ignore all the defaults and helpers and simply use the image as a stable base for libvirt installer development.