Oshinko Oshizushi

An operator that encapsulates the source-to-image process for creating Apache Spark based Deployments and Jobs within OpenShift.

Dependencies

This operator requires that you also have the Apache Spark Operator installed.

Quickstart

Prerequisites

1. A terminal shell with the OpenShift client tool (oc) installed.

2. An active login to OpenShift with cluster-admin role privileges. (this is required to install the custom resource definitions)

3. A copy of this repository.

Procedure

1. Install the custom resource definitions.

   oc apply -f deploy/crds
   

2. Install the operator.

   oc apply -f deploy
   

3. Create an application.

   oc apply -f examples/oshinkoapplication-sparkpi-python27.yaml
   

You should now see the operator creating an ImageStream, Build, and Deployment for the application. You will also see a Spark cluster being deployed.

Validation

To check that the application is working, you can run the following commands to create and expose an external route and then access the application for an estimate of the value for Pi. For a longer description of this application and its functionality, please see this tutorial.

1. Create a service

   oc create service clusterip sparkpi --tcp=8080
   

2. Expose the service

   oc expose svc/sparkpi
   

3. Curl the endpoint for a terrible estimate of Pi

   curl http://`oc get routes/sparkpi --template='{{.spec.host}}'`/sparkpi
   

If everything has worked you will see a message from the service returning an estimate for the value of Pi.