OpenShift Terraform

This project sets up an OpenShift Origin cluster (v3.9 currently) on a DigitalOcean infrastructure. The infrastructure creation and generation of inventory files is handled by Terraform. The goal of the project is to eventually add support for various cloud providers.

Overview

A lot of this setup is insipred by Get up and running with OpenShift on AWS. Here's a 10k ft view of how the setup looks like.

TODO: add diagram

Prerequisites

1. Terraform Download and install terraform for your OS.
2. Ansible
3. A DigitalOcean account with a ReadWrite Token generated.

Creating the Cluster

Provision the infrastructure

$ export TF_VAR_domain=<your-base-domain-name>
$ export DIGITALOCEAN_TOKEN=<do-token>
$ terraform init
$ terraform apply

Generate SSH keypair specific to this cluster. It is called tf and tf.pub by default. If you call it by a different name, change it in variables.tf.

$ terraform apply

This will generate 2 inventory files,

preinstall-inventory.cfg for installing the prerequisites in the Open Shift cluster.

inventory.cfg needed for the Ansible OpenShift installer.

Install prerequisite software

$ ansible-playbook -u root --private-key=~/.ssh/tf -i preinstall-inventory.cfg ./pre-install/playbook.yml

Install OpenShift cluster on provisioned infrastructure

Clone the playbook for v3.10.

$ git clone [email protected]:openshift/openshift-ansible.git --branch release-3.10 --depth 1
$ cd openshift-ansible

Install the verion of Ansible specified in the above repo. To do this, create a virtualenv, and run:

$ pip install -r requirements.txt

Prerequisite check.

$ ansible-playbook -i ../inventory.cfg --private-key=~/.ssh/tf playbooks/prerequisites.yml

Provision the cluster.

$ ansible-playbook -i ../inventory.cfg --private-key=~/.ssh/tf playbooks/deploy_cluster.yml

Post install steps

Login to master node using,

$ ssh -i ~/.ssh/tf [email protected]

And to the nodes using,

$ ssh -i ~/.ssh/tf [email protected]

Inside master node,

Create admin user.

$ htpasswd -cb /etc/origin/master/htpasswd admin <password>

Set up GlusterFS as the default storageclass.

$ kubectl get storageclass
NAME                PROVISIONER               AGE
glusterfs-storage   kubernetes.io/glusterfs   16m

$ oc patch storageclass glusterfs-storage -p '{"metadata":{"annotations":{"storageclass.beta.kubernetes.io/is-default-class":"true"}}}'

Deploy Drupal s2i image

Import the image into the openshift registry.

$ oc project openshift # all image streams should belong to this namespace
$ oc import-image lakshminp/openshift-drupal:v11 --confirm

Add the template either via UI or via cli.

Day 2 stuff

Adding new nodes

Destroying the Cluster

$ terraform destroy

CI

Contact

• email me
• lakshminp.com
• @lakshminp