Oracle Cloud Infrastructure Documentation

Creating the OpenShift Container Platform Cluster in OCI

Use these instructions to create the OpenShift Container Platform on OCI cluster and verify that the cluster is running.

Review the following topics and create the resources listed in them if you don't have them.

After creating the Compute service custom image, you are ready to provision compute nodes to create the cluster. OpenShift Container Platform nodes are either control plane nodes (also called "master nodes") or compute nodes (also called "worker nodes").

The first control plane node created for the cluster must be assigned the rendezvous IP address. This is the IP address found in the /rendezvousIP file created in your local working directory by the agent installer program. See Creating configuration files for installing a cluster on OCI in the Red Hat documentation for more information.

After you provision the first control plane node with the assigned rendezvous IP address, you can create the remaining nodes. When doing so, provision all control plane nodes before you begin provisioning worker nodes. The remaining nodes don't need assigned IP addresses. As you provision these nodes using the custom image, they are automatically discovered and incorporated into the cluster. After all nodes are provisioned, you verify that the cluster is running using the OpenShift Container Platform CLI.

  1. Navigate to the Compute service, then select Instances
  2. Select Create instance.
  3. Enter a Name for the instance.
  4. Select the compartment that the OpenShift cluster is being created in.
  5. In the Image and shape section, select Change image and choose My images.
  6. Select Custom images. Ensure the Compartment is set to the compartment where your custom image is stored.
  7. Select the checkbox beside the name of your custom image, then click Select image.
  8. Select Change shape.
  9. In the Browse all shapes panel, choose a shape and ensure that the Number of OCPUs value is 4 or more, and the Amount of Memory (GB) value is 16 or more.
  10. Click Select shape to continue.
  11. Choose the VCN that you specified when editing the oci-ccm.yml file. This VCN should also be referenced in the cidr value of the machineNetwork parameter in the install-config.yaml file.
  12. For the Subnet, choose Select existing subnet, then choose the private subnet for the VCN.
  13. First master node only: In the Primary VNIC IP addresses section, select Manually assign private IPv4 address. In the IPv4 address field, enter the rendezvous IP address. For extra master nodes, and for worker nodes, skip this step.
  14. In the Add SSH keys section, select Paste public keys and enter the public key from the SSH key pair discussed in Prerequisites.
  15. In the Boot volume section, select Specify a custom boot volume size and specify the volume size, and VPU. See Block Volume Performance for information on volume performance units (VPUs)
    • Control plane nodes (master nodes): Specify 500 GB for the volume size, and 60 for VPU.
    • Compute nodes (worker nodes): Specify 100 GB for the volume size, and 30 for VPU.
  16. Click Show advanced options to add a tag to the instance. Using the defined tag option, specify the openshift_tags namespace and the instance_role tag. Then assign either master (for control plane nodes) or worker as the tag value, depending on the node type.
  17. Select Create to provision the compute instance.
  18. Repeat steps 1 to 17 for the rest of the control plane nodes. Then use the same steps to provision the worker nodes. Optionally, you can create instance configurations for the control plane and worker nodes for rapid provisioning. See Creating an Instance Configuration for instructions on creating instance configurations.
  19. Use the instructions in Verifying that your Agent-based cluster installation runs on OCI in the Red Hat documentation to verify that the cluster is running. This step is performed in the OpenShift Container Platform CLI.