Manual Quick Start Guide

You need two VMs (LPARs): one bastion host and one SNO node. Both must have internet access and static IP addresses assigned before starting.

Required Access & Accounts

• Root access on the bastion host
• HMC access (for lpar_netboot) or SMS console access on the SNO LPAR
• Red Hat pull secret – download from cloud.redhat.com
• Internet access from both VMs (or a configured mirror registry)

Services Configured on Bastion

Sample Network Values Used in This Guide

Clone the ocp4-ai-power repository to your bastion host:

git clone https://github.com/ocp-power-automation/ocp4-ai-power.git
cd ocp4-ai-power

Generate SSH Key

Create the SSH key that will be used for OCP installation:

ssh-keygen -t rsa -b 2048 -N '' -C 'BASTION-SSHKEY' -f ~/.ssh/id_rsa

Save Pull Secret

Download your pull secret from cloud.redhat.com and save it on the bastion:

mkdir -p ~/.openshift
# Paste your pull secret content into this file
vi ~/.openshift/pull-secret

Setup Passwordless SSH to HMC

Configure passwordless access from the bastion to your HMC (required for lpar_netboot):

# Replace  with your HMC user and  with your HMC IP
ssh @
mkauthkeys -a "$(cat ~/.ssh/id_rsa.pub)"

All commands in this step are bundled in the prepare-bastion.sh script included in this repository. Run it directly on the bastion as root:

sudo bash prepare-bastion.sh

Manual steps – run the following on the bastion (supports RHEL 8/9 and CentOS 8/9):

# Detect OS and install Ansible
DISTRO=$(lsb_release -ds 2>/dev/null || cat /etc/*release 2>/dev/null | head -n1 || echo "")
OS_VERSION=$(lsb_release -rs 2>/dev/null || grep "VERSION_ID" /etc/*release 2>/dev/null \
  | awk -F= '{print $2}' | tr -d '"' || echo "")

if [[ "$DISTRO" != *CentOS* ]]; then
  # Red Hat
  if [[ $(cat /etc/redhat-release | sed 's/[^0-9.]*//g') > 8.5 ]]; then
    sudo subscription-manager repos --enable codeready-builder-for-rhel-9-ppc64le-rpms
    sudo yum install -y ansible-core
  else
    sudo subscription-manager repos --enable ansible-2.9-for-rhel-8-ppc64le-rpms
    sudo yum install -y ansible
  fi
else
  # CentOS
  if [[ $OS_VERSION != "8"* ]]; then
    sudo yum install -y https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm
    sudo yum install -y ansible-core
  else
    sudo yum install -y epel-release epel-next-release
    sudo yum config-manager --set-enabled powertools
    sudo yum install -y ansible
  fi
fi

# Install required Ansible collections
sudo ansible-galaxy collection install community.crypto --upgrade
sudo ansible-galaxy collection install community.general --upgrade
sudo ansible-galaxy collection install ansible.posix --upgrade
sudo ansible-galaxy collection install kubernetes.core --upgrade

# Install required system packages
sudo yum install -y wget jq git net-tools vim tar unzip python3 python3-pip \
  python3-jmespath coreos-installer grub2-tools-extra bind-utils

# Create PXE TFTP directory structure
sudo grub2-mknetdir --net-directory=/var/lib/tftpboot

Install dnsmasq to provide DNS, DHCP, and PXE services for the SNO node:

sudo yum install -y dnsmasq

Create /etc/dnsmasq.conf

##################################
# DNS
##################################
bogus-priv
enable-ra
bind-dynamic
no-hosts
expand-hosts

interface=        # e.g. env32
domain=sno.ocp.io
local=/sno.ocp.io/
address=/apps.sno.ocp.io/9.47.87.82  # SNO node IP
server=9.9.9.9

addn-hosts=/etc/dnsmasq.d/addnhosts

##################################
# DHCP
##################################
dhcp-ignore=tag:!known
dhcp-leasefile=/var/lib/dnsmasq/dnsmasq.leases

dhcp-range=9.47.87.82,static

dhcp-option=option:router,9.47.95.254
dhcp-option=option:netmask,255.255.240.0
dhcp-option=option:dns-server,9.47.87.83  # Bastion IP

dhcp-host=fa:b0:45:27:43:20,sno-82,9.47.87.82,infinite

###############################
# PXE
###############################
enable-tftp
tftp-root=/var/lib/tftpboot
dhcp-boot=boot/grub2/powerpc-ieee1275/core.elf

Create /etc/dnsmasq.d/addnhosts

9.47.87.82 sno-82 api api-int

Enable and Start dnsmasq

sudo systemctl enable --now dnsmasq
sudo systemctl status dnsmasq

Create the GRUB configuration file that tells the SNO node how to boot RHCOS and load the ignition config.

Create /var/lib/tftpboot/boot/grub2/grub.cfg

default=0
fallback=1
timeout=1

if [ ${net_default_mac} == fa:b0:45:27:43:20 ]; then
default=0
fallback=1
timeout=1
menuentry "CoreOS (BIOS)" {
   echo "Loading kernel"
   linux "/rhcos/kernel" ip=dhcp rd.neednet=1 \
     ignition.platform.id=metal ignition.firstboot \
     coreos.live.rootfs_url=http://9.47.87.83:8000/install/rootfs.img \
     ignition.config.url=http://9.47.87.83:8000/ignition/sno.ign

   echo "Loading initrd"
   initrd  "/rhcos/initramfs.img"
}
fi

Download the RHCOS kernel, initramfs, and rootfs images for ppc64le. These are served to the SNO node during PXE boot.

# Set the RHCOS base URL (adjust version to match your target OCP version)
export RHCOS_URL=https://mirror.openshift.com/pub/openshift-v4/ppc64le/dependencies/rhcos/4.12/latest/

# Create required directories
sudo mkdir -p /var/lib/tftpboot/rhcos
sudo mkdir -p /var/www/html/install
sudo mkdir -p /var/www/html/ignition

# Download PXE kernel and initramfs (served via TFTP)
cd /var/lib/tftpboot/rhcos
sudo wget ${RHCOS_URL}/rhcos-live-kernel-ppc64le -O kernel
sudo wget ${RHCOS_URL}/rhcos-live-initramfs.ppc64le.img -O initramfs.img

# Download rootfs image (served via HTTP)
cd /var/www/html/install
sudo wget ${RHCOS_URL}/rhcos-live-rootfs.ppc64le.img -O rootfs.img

Install and Configure httpd

Install Apache httpd to serve the rootfs and ignition files on port 8000:

sudo yum install -y httpd

# Configure httpd to listen on port 8000
sudo sed -i 's/^Listen 80$/Listen 8000/' /etc/httpd/conf/httpd.conf

# Fix SELinux context on web root
sudo restorecon -vR /var/www/html || true

sudo systemctl enable --now httpd

Download openshift-install

# Download the openshift-install binary for ppc64le (adjust version as needed)
wget https://mirror.openshift.com/pub/openshift-v4/ppc64le/clients/ocp/4.12.0/openshift-install-linux-4.12.0.tar.gz
tar xzvf openshift-install-linux-4.12.0.tar.gz

Create the Work Directory

mkdir -p ~/sno-work
cd ~/sno-work

Create ~/sno-work/install-config.yaml

apiVersion: v1
baseDomain:                   # e.g. ocp.io
compute:
- name: worker
  replicas: 0                        # No workers for SNO
controlPlane:
  name: master
  replicas: 1                        # Single control plane node
metadata:
  name:                # e.g. sno
networking:
  clusterNetwork:
  - cidr: 10.128.0.0/14
    hostPrefix: 23
  machineNetwork:
  - cidr:      # e.g. 9.47.80.0/20
  networkType: OVNKubernetes
  serviceNetwork:
  - 172.30.0.0/16
platform:
  none: {}
bootstrapInPlace:
  installationDisk:           # e.g. /dev/sda
pullSecret: ''
sshKey: |
  

Generate the Ignition File

# Generate the single-node ignition config
./openshift-install --dir=~/sno-work create single-node-ignition-config

# Copy the ignition file to the HTTP server directory
sudo cp ~/sno-work/single-node-ignition-config.ign /var/www/html/ignition/sno.ign
sudo restorecon -vR /var/www/html || true

curl http://9.47.87.83:8000/ignition/sno.ign | head -c 100

There are two ways to PXE-boot the SNO LPAR on PowerVM:

Option A – Using lpar_netboot on HMC (Recommended)

Run the following command from the HMC command line:

lpar_netboot -i -D -f -t ent \
  -m  \
  -s auto -d auto \
  -S  \
  -C  \
  -G  \
   default_profile 

Option B – Using SMS Console (Interactive)

Access the LPAR's SMS (System Management Services) console and manually select the network boot device. Refer to the HMC documentation for detailed SMS navigation steps.

After the SNO LPAR boots from PXE, monitor the installation from the bastion using openshift-install:

# Step 1: Wait for bootstrap to complete (typically 15-20 minutes)
cd ~/sno-work
./openshift-install wait-for bootstrap-complete --log-level=info

# Step 2: After bootstrap completes, wait for full installation
./openshift-install wait-for install-complete --log-level=info

Check Status with oc

You can also monitor progress using the OpenShift CLI:

# Set kubeconfig
export KUBECONFIG=~/sno-work/auth/kubeconfig

# Check node status
oc get nodes

# Check overall cluster version / installation progress
oc get clusterversion

# Check cluster operators (all should be Available=True when done)
oc get co

# Check pod status across all namespaces
oc get pod -A

After a successful installation, cluster credentials are stored in ~/sno-work/auth/:

• kubeconfig – kubeconfig file for CLI access
• kubeadmin-password – password for the kubeadmin user

CLI Access

# Set kubeconfig
export KUBECONFIG=~/sno-work/auth/kubeconfig

# Or login with kubeadmin credentials
oc login https://api.sno.ocp.io:6443 \
  -u kubeadmin \
  -p $(cat ~/sno-work/auth/kubeadmin-password)

# Verify cluster info
oc cluster-info
oc get nodes

Web Console

Open the following URL in your browser and login with user kubeadmin:

https://console-openshift-console.apps..
# Example: https://console-openshift-console.apps.sno.ocp.io

Download oc CLI

• Linux (ppc64le)
• Linux (x86_64)
• Mac OSX
• Windows