sbakiu / kubeflow-spark

Orchestrate Spark Jobs from Kubeflow Pipelines and poll for the status.

Kubeflow Spark

Orchestrate Spark Jobs using Kubeflow, a modern Machine Learning orchestration framework. Read related blog post.

Requirements

1. Kubernetes cluster (1.17+)
2. Kubeflow pipelines (1.7.0+)
3. Spark Operator (1.1.0+)
4. Python (3.6+)
5. kubectl
6. helm3

Getting started

Run make all to start everything and skip to step 6 or:

1. Start your local cluster

./scripts/start-minikube.sh

2. Install Kubeflow Pipelines

./scripts/install-kubeflow.sh

3. Install Spark Operator

./scripts/install-spark-operator.sh

4. Create Spark Service Account and add permissions

./scripts/add-spark-rbac.sh

5. Make Kubeflow UI reachable

• a. (Optional) Add Kubeflow UI Ingress

./scripts/add-kubeflow-ui-ingress.sh

• b. (Optional) Forward service port, e.g:

kubectl port-forward -n kubeflow svc/ml-pipeline-ui 8005:80

6. Create Kubeflow pipeline definition file

python kubeflow_pipeline.py

7. Navigate to the Pipelines UI and upload the newly created pipeline from file spark_job_pipeline.yaml

8. Trigger a pipeline run. Make sure to set spark-sa as Service Account for the execution.

9. Enjoy your orchestrated Spark job execution!