RedHat OpenShift Installation

Table of Contents

• Prerequisites
• Supported OpenShift Environments
• Installation

Introduction

This guide’s intention is to help a user deploy TIBCO ModelOps on RedHat OpenShift Platform, regardless of which platform OpenShift is installed on.

Prerequisites

1. Create a RedHat account
2. Ensure OpenShift is installed
3. Optional, In case of windows node is required, install opensift with hybrid network
4. Download and install Tools Kubectl and OC (OpenShift CLI)
5. Download and Install HELM Cli
6. Download and Install Tekton CLI (Optional)
7. Download and Install Lens (Optional)

Supported OpenShift Environments

ModelOps is supported on OpenShift running in these environments:

• vSphere (Datacenter)

ModelOps supports both installer-provisioned infrastructure (IPI) as well as user-provisioned infrastructure (UPI). Installer-provisioned infrastructure is recommended for a ModelOps installation.

To install an OpenShift cluster please use the official RedHat installation documents:

• vSphere (IPI & UPI)

Installation

1. Log in into OpenShift cluster, using any of the two ways below:

   1. Using the kubeconfig file:

      export KUBECONFIG = $path/kubeconfig
      

   2. Using OC for login:

      oc login \
          https://api.${clustername}.${domain}:6443 \
          -u ${cluster_username} \
          -p ${cluster_password}
      

2. Create a namespace

    kubectl create namespace ${namespace}
   

3. Add the secrets

   kubectl create secret generic elasticsearch-es-elastic-user \
       --from-literal=${elastic_username}=${elastic_password} \
       --namespace ${namespace} --dry-run=client --output=yaml 2>/dev/null > secret.yaml
   kubectl apply --filename secret.yaml
   
   kubectl create secret generic git-server \
       --from-literal=${git_username}=${git_password} \
       --namespace ${namespace}
   
   kubectl create secret generic nexus-server \
       --from-literal=${nexus_username}=${nexus_password} \
       --namespace ${namespace}
   
   kubectl create secret generic modelops-server \
       --from-literal=${modelops_username}=${modelops_password} \
       --namespace ${namespace}
   
   kubectl create secret generic scoring-admin \
       --from-literal=${scoring_username}=${scoring_password} \
       --namespace ${namespace}
   

4. Install Helm charts

   helm upgrade \
       --install installer ${home}/helm-charts/kubernetes-installer-1.0.2.tgz \
       --set cloud=openshift \
       --set openshift.ingressDomain=$(kubectl get ingresses.config/cluster -o jsonpath={.spec.domain}) \
       --namespace ${namespace}
   

5. Copy the shipped Maven artifacts to the Maven repository Pod

   kubectl cp \
       ${path}/modelops-repo-1.2.0-mavenrepo.zip mavenrepo-0:/tmp \
       --namespace ${namespace}
   

   This command takes some time to run, and gives no output.

6. To monitor installation logs, use:

   tkn pipelinerun logs bootstrap --follow --namespace ${namespace}
   

7. The installation is completed when all tasks display Succeeded

    $ tkn taskrun list --namespace ${namespace}
   

8. Capturing ModelOps UI url

   1. Log into OpenShift console

      To get the console URL run

       $ oc whoami --show-console
      

      Navigate to the URL and pass ${cluster_username} and ${cluster_password}

   2. Select developer mode

      

   3. Navigate to Topology and Click on ModelOps server