Introduction
This project is to provide one-click experience for people to create Cloud and Kubernetes environment to run their Data Analytics workload
like Apache Spark.
For example, "punch install SparkOnK8s" will create an AWS EKS cluster and a Spark REST Service. Then you could use curl or command line tool
to submit Spark application.
How to build (on MacBook)
The following command will create dist
folder and dist.zip
file for Punch.
make release
Go to dist
folder, then check User Guide to see how to run punch command.
Quick Start
You could build this project (make release
) and use punch
to deploy Spark on Minikube, and run Spark application for a quick try.
You could also use Homebrew to install punch
and have a try.
See Quick Start Guide for details.
TODO
- Attach tag (e.g. punch-topology=xxx) to AWS resources created by punch
- Mask password value in helm output (e.g. --set apiGateway.userPassword=xxx)
- Remove unnecessary argument like "--set apiUserPassword=password1" in "punch uninstall" command
- Allow patch topology like --patch foo.field1=value1
- Allow set values by file like --values values.yaml
- Return HTTP 404 when sparkcli getting a non-existing application
- Get application error message from Spark Operator