fcarrus / ocp4_deployment

OpenShift 4 Deployment Automation

Ansible Collection - fcarrus.ocp4_deployment

OpenShift 4 Provisioning and Configuration using Ansible.

Runs in a container, no need for additional libraries or software, aside from podman.

What it does

• Helps with the manual tasks of creating the needed VMs for OpenShift nodes
• Configures several aspects of an OpenShift cluster (i.e. authentication, logging, monitoring, ODF storage, and more...)
• Can scale up the cluster, creating additional nodes.
• Strives to be idempotent at any phase.
• Can configure external services (NFS shares on an existent machine, ATM).

Where it can be used

Only VMware at the moment. More to come...

How to use it

The automation runs in a container, so you'll need a RHEL8-like machine with podman installed. You don't need root privileges to run it.

Pull the container:

$ podman pull quay.io/fcarrus/ocp4-deployment-automation:latest

Create a directory where to keep the automation inventory and all the installation-related files, i.e.:

$ mkdir /home/myuser/mycluster

Use the inventory.yaml.example file in this repository to define your cluster's configuration (nodes, post-deploy config, etc.). Documentation on variables is in the example file.

Place the file in the directory created above and name it inventory.yaml.

Use the run.sh script in this repository as a template for your own script. You should edit the MYDIRECTORY variable and assign it the directory you created, i.e.:

MYDIRECTORY="/home/myuser/mycluster/"

Run the automation's check playbook. It validates some of the variables in the inventory, so it can fail early before starting the deployment.

$ ./run.sh check

Run the automation's prepare playbook. It creates the install-config.yaml, manifests and ignition files. They are all placed in the directory you created.

$ ./run.sh prepare

Run the automation's deploy playbook. It creates the CoreOS template image and the node VMs, then waits for the cluster to be up&running.

$ ./run.sh deploy

Run the automation's configure playbook. It uses the kubeconfig file to configure the cluster as specified in the inventory file.

$ ./run.sh configure

The configure playbook can be run for single parts of configuration at a time, making use of Ansible tags, i.e.:

$ ./run.sh configure -t node-labels,ingress,logging

All the tags available for the configure playbook are:

• api-certificate : Replace the API endpoint's SSL certificate.
• ca-bundle: Inject certification authorities in the trusted bundle.
• defns: Set up the default namespace template adding NetworkPolicies.
• idp: Set up OAuth Identity Provider(s).
• image-registry: Configure the Image Registry.
• ingress: Configure the Ingress Controller.
• logging: Install the Logging stack and configure it.
• manifests: Add objects to OpenShift in post-deploy phase.
• monitoring: Configure the monitoring stack.
• node-labels: Label nodes.
• node-taints: Apply node taints.
• odf: Install ODF and create a StorageCluster object.
• proxy: Configure cluster-wide proxy settings.
• registry-sources: Configure additional registry sources.
• rhacm: Install RHACM and configure the cluster as Hub.
• selfprov: Disable the self-provisioning feature.
• update-cap: Updates the cluster auto approver with additional nodes after scaling up.

Other uses

There is a playbook for configuring an external RHEL-like server with NFS shares (see documentation in the inventory example file):

$ ./run.sh external_nfs

References

VMware vCenter user : required account privileges