This contains scripts for running various types of clusters on CU Boulder computing resources. This is primarily targeted at running clusters on the educational and research.
If they work on other platforms that is unintentional/unsupported, but possible with changes to the Packer scripts.
The primary tool used to provision clusters is Packer.
Packer is a tool to create images with the correct software provisioned. Packer works by launching an instance in a compute environment, then running a set of provisioning steps to bring the instance to the desired state. Packer then takes a snapshot of the instance to make an image, destroys the instance, and publishes the image to the computing environment.
Currently, this repository contains
- Scripts for running a Spark cluster within
spark/ - Scripts for running a Docker Swarm cluster within
docker/