In this article, I will explain how we can easily deploy a traditional WebSphere application on Openshift Container Platform (OCP) on Google Cloud (GCP) and WebSphere application server. Openshift Container Platform (OCP) is a hybrid cloud enterprise Kubernetes application platform.

We will start by installing the Openshift Container Platform in Google Cloud using a simple single command. Following that, we will detail the steps on how a traditional WebSphere application can be deployed on a containerized WebSphere Application Server (WAS) running on OCP with zero code changes.

The steps needed to install Openshift Container Platform in Google Cloud Platform (GCP) and how  to deploy the “Hitcount” application on the WebSphere Application Server are listed below

Preparatory Steps:

  • Download OpenShift installer binary and OpenShift Command line client (oc)
  • Pull secret to install OCP
  • GCP - Google Cloud account, CLI tool, and a billing account
  • DNS domain
  • SSH key

Preparatory steps for openshift-install:

  • Openshift binaries and pull secrets can be downloaded using this process. You can sign in to and, select the self-managed option to get both the pull secret code and the required binaries for install and openshift CLI tools.

Obtain the pull secret from

  • Specific versions of openshift-install binary and tools

To install a specific version of OCP, you will need the corresponding installer programs that can be executed from a laptop or a server. Use this link to choose the version; for example,  to install container version 4.7.5, download executables from the following 

Download the file openshift-install-<os name>.... and unpackage the file to get the installer file “openshift-install” . Download and extract the install program for your operating system and place the file in the directory where you will store the installation configuration files.

For example: openshift-install-mac-4.7.5.tar.gz is the installer for the 4.7.5 to run on Mac OS

  • Download the OpenShift command-line tools and add them to your PATH.

Download the file openshift-client-<os name>.... And unpackage the file to get oc and kubectl command line utilities.

For example: openshift-client-mac-4.7.5.tar.gz contains the client tools oc, kubectl

Preparatory steps for GCP:

  • Google cloud account - to get access to a Google GCP Account
  • Install Google cloud SDK gcloud CLI tool - google-cloud-sdk from
  • Create project in GCP from Google cloud console - eg. ocp4-on-gcloud-rh001-
  • Initialize gcloud tool by running “gcloud init” setting the project name, compute region/zone eg. us-central1-a
  • We need to enable the required services APIs for the project to allow the creation of required cloud resources:
  • Since we will be using Installer provisioned infrastructure (IPI) to install OpenShift, you need to make sure  there is enough availability of the following resources for the region. In Google console IAM & Admin menu option:
  • DNS Domain will be required to host the OCP container.
  • SSH key readiness - ssh-keygen -t ed25519 -N '' -f ~/.ssh/rsa_key_1

Or from CLI tool

export gcp_project = ocp4-on-gcloud-rh001
gcloud projects create ${gcp_project}
gcloud config set project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}
gcloud services enable --project ${gcp_project}

Link billing to the project:

gcloud alpha billing projects link ${gcp_project} --billing-account $(gcloud alpha billing account list | tail -l | awk ‘{print $1}’) 

Note: When it is created in the console it will be automatically associated to the billing account

Verify GCP account is tied to a billing account by checking the Billing menu on the GCP console.

Service account is responsible for creating resources in google, so create a service account and attach it as a owner to the project:

export gcp_sa=ocp4-sa
gcloud iam service-accounts create ${gcp_sa}
gcloud projects add-iam-policy-binding ${gcp_project} --member "serviceAccount:${gcp_sa}@${gcp_project}" --role "roles/owner"
gcloud iam service-accounts keys create ${HOME}/.gcp/osServiceAccount.json --iam-account ${gcp_sa}@${gcp_project}



Edit quota and make a request for the Persistent Disk SSD (GB) for the region (for example, . us-central-1) from 500 GB to 950 GB. Similarly request CPUs increase from 24 to 32. Technically It could take a few hours, but we have noticed the requests get fulfilled in a few minutes.

DNS Domain:

This can be bought from the GCP itself or needs to be configured in GCP

(for example, or it can be subdomain like This can be bought directly from Google or from another DNS provider.

SSH key:

Have a key pair ready that can be used for debugging or recovery purposes

This will be used during installation, under a directory named .ssh from the main installation directory.

Openshift Installation:

We will be using Installer provisioned infrastructure option(IPI) with the following simple command: openshift-install create cluster --dir=./ocpproject/

You can choose any directory name instead of ocpproject.

Platform: choose gcp

Project ID

bradhak@abradhak-mac ~ % openshift-install create cluster --dir=./ocpproject/
? SSH Public Key /Users/abradhak/.ssh/
? Platform gcp
INFO Credentials loaded from file "/Users/abradhak/.gcp/osServiceAccount.json"
? Project ID ocp4-on-gcloud-rh001 (ocp4-on-gcloud-rh001)
? Region us-central1
? Base Domain
? Cluster Name ocp4test
? Pull Secret [? for help] ***************************************************************

Note: Output will look something like this

INFO Creating infrastructure resources...     ****************************************INFO Waiting up to 20m0s for the Kubernetes API at ********************
INFO Waiting up to 30m0s for bootstrapping to complete...
**INFO Destroying the bootstrap resources...    
********INFO Waiting up to 40m0s for the cluster at to initialize...
**INFO Waiting up to 10m0s for the openshift-console route to be created... ***INFO Install complete!              
***INFO To access the cluster as the system:admin user when using 'oc', run 'export KUBECONFIG=/Users/abradhak/ocpproject/auth/kubeconfig'
**INFO Access the OpenShift web-console here: ****INFO Login to the console with user: "kubeadmin", and password: "Pev…...-Gens4" ******************************************************************************INFO Time elapsed: 43m4s  

You can try some sample oc commands following the installation completion by executing

export KUBECONFIG=/Users/abradhak/ocpproject/auth/kubeconfig
oc get nodes


Application Deployment:

Log in to the openshift web console as seen in the output above ( ) using kubeadmin and the password obtained from the installation logs

You can retrieve it by executing the below:

        cat .openshift_install.log

Create a project in Openshift from the developer’s mode:

We will be using the Docker build strategy to build the application image.

Let us review the GitHub repo (

It has a Docker file, has the application’s .ear file and the WebSphere server’s configuration files (analyzed and generated using IBM Transformation advisor).


# Install WebSphere application server from universal base image
FROM ibmcom/WebSphere-traditional:
# Copy the application package ear file to the container working directory
COPY --chown=was:root defaultapplication.ear /work/config/defaultapplication.ear
# put app and scripts and properties in /work/config
# put external library (e.g db driver) in /work/config/lib
COPY --chown=was:root ./src/config /work/config
COPY --chown=was:root ./lib /work/config/lib
# Start the application server
RUN /work/

The main point to be noted is there is “zero code change” required to migrate an existing WebSphere application to a container based deployment.

Building and Deploying Hitcount Application

Select the import from Docker file option and populate the following fields:

Git Repo URL - 

Use an application name and name as you may prefer (for example, ta-ark-demo-app).

Select the resource type to be generated as “Deployment Config,” and click on the Create button. This will start building and deploying the application to the OCP: 

You can see the build logs under Builds -> Build details as seen below:

Now, let us review the topology view, as you can see the build success denoted by the green arrow.

As seen in the below screen capture, modify the OpenShift Service’s port, target port, and name to use port 9080 instead of the default 8080. Click on the service and YAML tab to get to the below screen:

The application can be tested using the route location and adding the application name (hitcount)

Select the inputs as specified to test the application a few times by using both Google Chrome and an alternate browser (Safari or Firefox) to verify the application.


Rehosting a hitcount application on to OpenShift without making any code changes was demonstrated above. This is an indicator that other traditional WebSphere apps can be rehosted on OpenShift with zero application changes. Even bigger and complex IBM WebSphere applications can follow a similar approach to rehost, re-platform, and containerize in OpenShift on GCP and on other infrastructure easily and with zero application changes. Take a few minutes to check out the associated short videos in the reference section below.


OpenShift Install:

Reference videos on multiclouds and on-premises for WAS application deployment:

We also created a series of videos demonstrating the deployment of the traditional WebSphere JEE application to a OCP cluster in Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platfom(GCP), and on-premises VMWare VSphere.
Code references:


How-tos, Java, IBM, workloads, GCP, WebSphere

< Back to the blog