Single Node OpenShift is a new offering that also works great for development, although it does require some planning and preparation before using it for the first time. You will need to provide 8 vCPU, 32GB of RAM, and 120GB of storage at a minimum.

  • Decide where you want to run SNO after reviewing the minimum requiremnents
  • Configuring a static lease is helpful but not required
  • Configuring DNS entries is preferable, but again optional

Minimum Requirements:

  • vCPU: 8
  • RAM: 32GB
  • Storage: 120GB

In the example below I used a KVM virtual machine with a static lease and DNS entries created in dnsmasq.


Cluster Details

On this page we will fill in some basic information

  • Give the cluster a name
  • Choose the base domain
  • Check Install single node OpenShift (SNO)
  • Choose version 4.9.0
  • Click Next

Host Discovery

On this page we will set up a discovery image for our cluster to boot the VM off of.

  • Click Generate Discovery Image
  • Enter an ssh key.
  • Click Generate Discovery ISO
  • Use one of the provided options to download the ISO

In the next section we will be configuring a VM and booting it, but do not close this page. We will be returning to it.

Virtual Machine Creation

We must now configure a VM to boot off of the ISO we downloaded. In the following steps I will use Virtual Machine Manager to create a KVM VM, but other virtualization options are likely also acceptable.

  • Start Virtual Machine Manager
  • Click File>New Virtual Machine
  • Leave the default `Local install media (ISO Image or CDROM)
  • Click Forward
  • Browse to the ISO you downloaded
  • Uncheck Automatically detect from installation media / source
  • Search for and select Red Hat Enterprise Linux 9.0 and select it
  • Click Forward
  • Set the Memory to 32GiB or greater
  • Set teh CPUs to 8 or greater
  • Click Forward
  • Set Storage to 120 GiB or greater
  • Click Forward
  • Update the VM name if desired
  • Check Customize configuration before install
  • Choose the networking option you want to use
  • Click Finish

Since we checked Customize configuration before install a new window will pop up.

If you would like to set a static lease review the MAC address assigned to the NIC now and do so. The means of doing this can vary drastically by environment, so I will not attempt to cover the process. Note that you should not name the host the same as your cluster. For example, if you configured your cluster to be the node should not be configured as, but would be acceptable.

When done click Begin Installation and the VM should boot from the ISO.

Finishing Discovery

Return to the assisted installer discovery page.

After a moment the VM you booted from the ISO should be discovered. If you set a static lease with a DNS name for the system it should appear with that name, otherwise you will likely see localhost. If you see localhost change the name to something else.

If you hit the drop down you may see NTP Status not yet ready. I find that the process is smoother if I wait until this is marked Ready, which may take a moment or two, though no action is required.

When done, click Next.


  • Choose your subnet in the Available Subnets drop down
  • Click Next

Review and create

  • If everything looks acceptable click Install cluster

As the installation proceeds you will be provided with a kubeconfig for the new cluster, and once it completes you will be provided with a password for the kubeadmin user.

Configure DNS

You may either configure DNS with two entries, or edit /etc/hosts on your system to access the new cluster. It is important that you do not create a wildcard address for the cluster domain. Instead create an entry for api.cluster-domain.base-domain, and a wildcard for apps.cluster-domain.base-domain.

In dnsmasq an example of this would look like:,

For bind:      A
*   A

Alternatively, modify /etc/hosts:

Optional SSH config

From time to time I have reprovisioned my cluster. Since this changes the SSH host keys for the cluster I have opted to not store them.

In my .ssh/config I have an entry for the node:

  StrictHostKeyChecking no
  UserKnownHostsFile /dev/null
  User core

PVs and Storageclass

For testing stateful migrations with MTC it is helpful to have a storageclass and associated PVs available. As SNO does not configure any by default you will need to do so yourself. With the ssh configuration above, and my ssh key unlocked with my ssh agent, I run the following script:


export KUBECONFIG=~/Documents/sno-kubeconfig

ssh -i ~/.ssh/id_rsa_ocp "sudo /bin/bash -c 'mkdir -p /srv/openshift/pv-{0..99} ; chmod -R 777 /srv/openshift ; chcon -R -t svirt_sandbox_file_t /srv/openshift'"

for i in {0..99}; do
  oc create -f - <<EOF
apiVersion: v1
kind: PersistentVolume
  name: pv-$i
    type: local
  storageClassName: manual
    storage: 10Gi
  - ReadWriteOnce
  - ReadWriteMany
  persistentVolumeReclaimPolicy: Recycle
    path: "/srv/openshift/pv-$i"

oc create -f - <<EOF
kind: StorageClass
  name: manual
  annotations: 'true'
volumeBindingMode: WaitForFirstConsumer

Shutting down, Rebooting, and Upgrades

Shutting down and rebooting your cluster should not present a problem.

Upgrades are supported as of SNO version 4.9.0. At the time of this blog the fast channel has an updated available up to version 4.9.5.

If my cluster was off for long periods of time the console would not come back up with version 4.9.0. So far I have not experienced this with 4.9.5, so I’d recommend upgrading after the initial install is complete.