Generating the discovery ISO manually

Installing OKD on a single node requires a discovery ISO, which you can generate with the following procedure.

  1. Download the OKD client (oc) and make it available for use by entering the following command:

    $ curl -k https://mirror.openshift.com/pub/openshift-v4/clients/ocp/latest/openshift-client-linux.tar.gz > oc.tar.gz
    $ tar zxf oc.tar.gz
    $ chmod +x oc
  2. Set the OKD version:

    $ OCP_VERSION=<ocp_version> (1)
    1 Replace <ocp_version> with the current version. For example. latest-4.10
  3. Download the OKD installer and make it available for use by entering the following commands:

    $ curl -k https://mirror.openshift.com/pub/openshift-v4/clients/ocp/$OCP_VERSION/openshift-install-linux.tar.gz > openshift-install-linux.tar.gz
    $ tar zxvf openshift-install-linux.tar.gz
    $ chmod +x openshift-install
  4. Retrieve the FCOS ISO URL:

    $ ISO_URL=$(./openshift-install coreos print-stream-json | grep location | grep x86_64 | grep iso | cut -d\" -f4)
  5. Download the FCOS ISO:

    $ curl -L $ISO_URL > rhcos-live.x86_64.iso
  6. Prepare the install-config.yaml file:

    apiVersion: v1
    baseDomain: <domain> (1)
    - name: worker
      replicas: 0 (2)
      name: master
      replicas: 1 (3)
      name: <name> (4)
      networkType: OVNKubernetes
      - cidr: <IP_address>/<prefix> (5)
        hostPrefix: <prefix> (6)
      - <IP_address>/<prefix> (7)
      none: {}
      installationDisk: <path_to_install_drive> (8)
    pullSecret: '<pull_secret>' (9)
    sshKey: |
      <ssh_key> (10)
    1 Add the cluster domain name.
    2 Set the compute replicas to 0. This makes the control plane node schedulable.
    3 Set the controlPlane replicas to 1. In conjunction with the previous compute setting, this setting ensures the cluster runs on a single node.
    4 Set the metadata name to the cluster name.
    5 Set the clusterNetwork CIDR.
    6 Set the clusterNetwork host prefix. Pods receive their IP addresses from this pool.
    7 Set the serviceNetwork CIDR. Services receive their IP addresses from this pool.
    8 Set the path to the installation disk drive.
    9 Copy the pull secret from the Red Hat OpenShift Cluster Manager. In step 1, click Download pull secret and add the contents to this configuration setting.
    10 Add the public SSH key from the administration host so that you can log in to the cluster after installation.
  7. Generate OKD assets:

    $ mkdir ocp
    $ cp install-config.yaml ocp
    $ ./openshift-install --dir=ocp create single-node-ignition-config
  8. Embed the ignition data into the FCOS ISO:

    $ alias coreos-installer='podman run --privileged --pull always --rm \
            -v /dev:/dev -v /run/udev:/run/udev -v $PWD:/data \
            -w /data quay.io/coreos/coreos-installer:release'
    $ cp ocp/bootstrap-in-place-for-live-iso.ign iso.ign
    $ coreos-installer iso ignition embed -fi iso.ign rhcos-live.x86_64.iso

Installing with USB media

Installing with USB media involves creating a bootable USB drive with the discovery ISO on the administration node. Booting the server with the USB drive prepares the node for a single node installation.

  1. On the administration node, insert a USB drive into a USB port.

  2. Create a bootable USB drive:

    # dd if=<path-to-iso> of=<path/to/usb> status=progress

    For example:

    # dd if=discovery_image_sno.iso of=/dev/sdb status=progress

    After the ISO is copied to the USB drive, you can use the USB drive to install OKD.

  3. On the server, insert the USB drive into a USB port.

  4. Reboot the server and enter the BIOS settings upon reboot.

  5. Change boot drive order to make the USB drive boot first.

  6. Save and exit the BIOS settings. The server will boot with the discovery image.

Monitoring the installation manually

If you created the ISO manually, use this procedure to monitor the installation.

  1. Monitor the installation:

    $ ./openshift-install --dir=ocp wait-for install-complete

    The server will restart several times while deploying the control plane.

  2. Optional: After the installation is complete, check the environment:

    $ export KUBECONFIG=ocp/auth/kubeconfig
    $ oc get nodes
    $ oc get clusterversion