Easily Provision and Manage Clusters at Massive Scale
“Warewulf 4 is the HPC cluster manager I always wanted. Its OCI integration streamlines and simplifies the image creation process, making it trivial to build customized cluster environments on the foundation of the growing container ecosystem.”
— Jonathon Anderson
It’s Your Work. Dramatically Transformed.
Whether you’re managing tens of nodes or tens of thousands, you want the simplest way to provision the operating system, a single point of control, and ease of configuration.
First Reality-Based Rule: Don’t ask a Systems Administrator whether they’ve ever felt overwhelmed unless you’re prepared to hear the answer. Feeling overwhelmed (at some point) just comes with the turf. Everybody’s got their stories.
So it was for Gregory Kurtzer when he was running Linux clusters at Lawrence Berkeley National Laboratory for the Department of Energy back in 2001. Based on the reality of his direct experience—that feeling of being overwhelmed—here’s what he knew he needed: Something completely different from existing cluster toolkits. He needed a simple, stateless way of provisioning and administering operating systems and configurations for clusters with lots of nodes. The goal was to dramatically transform an important part of his work, so, you know, he could feel overwhelmed by something else. Because there’s always something else, right?
That’s why Kurtzer founded Warewulf: a cluster provisioning solution that builds on a well-established stateless provisioning paradigm and incorporates enterprise-grade containerization workflows to make it easy to manage clustered hardware at scale.
Isn’t Just a Marketing Word.
Warewulf is a cluster management and provisioning platform for all types of clusters, and is designed to produce secure, scalable, turnkey cluster deployments that are both flexible and easy to use. From small, researcher-run clusters all the way to HPC clusters with tens of thousands of nodes, Warewulf scales both up and down to be easy to deploy, administer, debug, and secure in any clustering environment.
You Know, Like the Internet?
The Warewulf project pioneered the concept of stateless provisioning in High Performance Computing (HPC), setting the standard for large-scale cluster deployment. After many years of development, Warewulf transformed from a basic provisioning system into the leading cluster management platform for all HPC environments. (Just a side note: Technically speaking, Warewulf is not an HPC system. It solves the problem of slave node management in a way that is not limited to use in HPC-specific environments.)
Its stateless paradigm makes Warewulf highly scalable and easy to deploy, administer, debug, and secure in any clustering environment.
Warewulf Is Not Just Another Face in the Crowd
Easy to change and customize for any clustering use case
Minimal requirements and easy to get started
Warewulf enables HPC without getting in the way
From tens to thousands of nodes
Supports SELinux, per-node asset keys, and access controls
Stateless provisioning of entire clusters. Hard drives are optional
You Can Count On the Warewulf-Keepers.
Tell us a bit about yourself and your goals for Warewulf. We’ll get back to you promptly. Possibly with a silver bullet.