Step-by-step Deploy of Single Node OpenShift (SNO)

Red Hat recently released a technology preview of an assisted installer for single node OpenShift (SNO). A Red Hat blog post about this is available here.

As stated in the blog, more and more organizations are looking to deploy and manage containers at the edge, and that can introduce challenges related to space, networking, and other limitations.

On the topic of space, and edge, the International Space Station is actually running a single node OpenShift instance for DNA sequencing. More about that in this article.

So, let’s get started. First of all, you need to decide where to install the node. Cloud, virtual, or physical machine? All is supported, as long as you can boot the machine from the generated image. The following hardware requirements must be considered:

Basic install without virtualization:

CPU: 8 cores (or 4 cores with hyperthreading enabled)


HDD: 120GB

If you want to install with virtualization, CPU requirements are 12 cores/threads on bare metal.

And, of course, internet access, since SNO needs to communicate with our SaaS service.

With that part sorted, logon to with your Red Hat account. If you do not have one yet, you will be able to create one in the login page.

After logging in, select OpenShift as shown here:

Then select to “Create Cluster,” choose “Datacenter” as target, and “Assisted Installer” should show up like in this screen capture:

This will start the wizard, where you need to enter:

  • A cluster name
  • A domain name
  • And check boxes for SNO and that you understand the limitations of a technology preview.

Click “Next” when done.

If you have enough CPU cores (12), deploy on bare metal, and want virtualization integrated, select the checkbox for “Install OpenShift Virtualization.”

Then press “Generate Discovery ISO”:

In that dialogue, you select preferred media, USB or CD, and provide your public SSH key. For my environment, I just use the /.ssh/ key on my desktop computer.

After pressing “Generate Discovery ISO” you will get links to download the image.

In my scenario, I copied the ISO to a suitable datastore, and mounted it in a virtual machine and booted the machine from the ISO. For bare metal, you can either use the same ISO to boot via remote management solutions, or just generate a bootable USB-stick with for instance BalenaEtcher or similar.

After the target has booted successfully, it will show up in your browser as “localhost”. Give it a desired hostname like this:

Next step is networking. I am  doing this really simple, so I just select the desired available subnet in the pull down menu. You can enter advanced settings if you want to customize more related to networking.

I also select (default) to reuse my SSH key from the discovery host.

Clicking “Next” will show a summary like this:

Press “Install cluster” to start the install:

Depending on hardware and connectivity, the install will take around 45 minutes. When done, you will see a status like this:

Two things to remember before trying the URL and login:

Do you have a working DNS resolving for the host?

In my homelab, I just run a simple BIND server for that, so I added this, as described under “Not able to access the Web Console?” Remember to replace with your own domain/IP.                A
*        A
ocp1node                                A

An even simpler solution is to add this to your local hosts file:

Now we are ready to check out your install.

Click the “copy to clipboard” to the right of the password.

Click the Web Console URL. (Accept the SSL-warnings)

Username should be populated with “kubeadmin” paste your password and select “Log in”

… and tap yourself on the shoulder for job well done if you are greeted with this dashboard:

Now you can start to explore and play with OpenShift in a pretty fast and simple way.

For those that want even simpler ways to have a look at OpenShift, there are more options.

Developer Sandbox for OpenShift

Get started in seconds with OpenShift as a Service. It is totally free, but pods will not live past 12 hours, so this is not for persistent testing and evaluation.

Link to sign up and test:

Code Ready Containers:

If you have a laptop/desktop or preferably a workstation, you can also run Code Ready Containers (CRC), a local desktop version of OpenShift tailored to that scenario. Not all the functions and features as the “real thing,” but as a developer you will get a taste of UI/TUI possibilities. Warning: On a “normal” laptop, performance will not be “excellent.” SNO might be a better option.

In addition to the above mentioned ways to access OpenShift, you can also have a look at Microsoft Azure, Amazon AWS, and IBM Cloud OpenShift offerings to get started in the public cloud, quick and easy.

Red Hat OpenShift is supported by most public cloud vendors globally, by most major service providers, and of course on a vast array of platforms and hardware on premise. Or in space.


How-tos, Installation, OpenShift 4.8, single node

< Back to the blog