Welcome to elasticluster’s documentation!

Introduction

ElastiCluster aims to provide a user-friendly command line tool to create, manage and setup computing clusters hosted on cloud infrastructures (like Amazon’s Elastic Compute Cloud EC2, Google Compute Engine, or a private OpenStack cloud). Its main goal is to get a private cluster up and running with just a few commands; this video demoes ElastiCluster setting up a computational batch-queueing cluster.

Complete documentation for ElastiCluster is available on the Read The Docs website. General discussion over ElastiCluster’s usage, features, and bugs takes place on the elasticluster@googlegroups.com mailing-list (only subscribers can post).

The ElastiCluster project is an effort of the Services and Support for Science IT (S3IT) unit at the University of Zurich, licensed under the GNU General Public License version 3.

Features

ElastiCluster is in active development, and offers the following features at the moment:

  • INI-style configuration file to define cluster templates
  • Can start and manage multiple independent clusters at the same time * Automated setup of:
  • Growing and shrinking a running cluster.

ElastiCluster is currently in active development: please use the GitHub issue tracker to file enhancement requests and ideas, or the mailing-list for discussion.

We appreciate pull requests for new features and enhancements. Please use the master branch as starting point.

Overview

The architecture of elasticluster is quite simple: the configuration file ~/.elasticluster/config defines a set of cluster configurations and information on how to access a specific cloud service (including access id and secret keys).

Using the command line or a simple API, you can start a cluster (possibly overriding some of the default values, like the number of nodes you want to fire up) and configure it:

  • ElastiCluster connects to the cloud provider indicated in the cluster configuration file, starts virtual machines, and waits until they are accessible via ssh.
  • After all the VMs are up and running, ElastiCluster runs Ansible to configure the cluster.

Upon resize of the cluster [1], new virtual machines will be created and again Ansible will run on all the VMs, in order to properly add the new hosts to the cluster.

ElastiCluster commands export and import allow moving a running cluster’s definition and status data from one machine to the other, to allow controlling the same cluster from different places.

[1]Currently, only growing a cluster is fully supported; shrinking a loaded cluster may remove nodes with running jobs and cause malfunctionings. See the remove-node command for a safer, albeit more low-level, way of shrinking clusters.

Table of Contents

Indices and tables