SUSE Rancher and Hammerspace

How to Simplify and Accelerate the Delivery of Kubernetes Applications and Data

 

Introduction

Kubernetes has freed applications from the limitations of underlying server technologies, starting a revolution in the way applications are created, tested, and delivered to end-users. Kubernetes accelerates the production pipeline, drastically improving time to market. It is the perfect fit for a world increasingly dependent on agile development. Consequently, these higher demands are also putting pressure on IT departments to adopt similarly agile operating procedures. This marriage of agile Development practices and IT Operations is aptly named DevOps. If you are familiar with David Kim’s book, The Phoenix Project, you are already aware that DevOps is the adoption of Kanban process engineering practices, a key part of agile software development. DevOps shifts the primary focus away from managing the minutiae of technology towards managing optimal business outcomes, such as a smoothly running CI/CD pipeline. The outcome is higher efficiency, productivity and profitability.

 

SUSE Rancher

Founded in 1992, SUSE is the world’s first provider of an Enterprise Linux distribution.SUSE is also the largest independent open-source software company in the world. SUSE Rancher centrally manages your on-premises or hybrid clusters to ensure security and accelerate transformation. It is a product built to address the needs of both the DevOps teams deploying applications with Kubernetes and the IT staff responsible for delivering enterprise-critical services. SUSE Rancher supports any CNCF-certified Kubernetes distribution.

 

Hammerspace

Hammerspace is the world’s first storageless data solution. The architecture is an abstraction from infrastructure that provides a more powerful and scalable way of managing data across any infrastructure and delivering data services with multiprotocol access to users at any location. The Hammerspace data orchestration platform delivers storageless data for the automated deployment, scaling, management, and declarative control of data across disparate storage systems on any infrastructure.

 

Use Helm to Provision Storage

Helm makes provisioning Hammerspace storage for Rancher an easy and straightforward task. It reduces complexity and removes error-prone manual configuration steps through automation. All you have to do is go to Rancher Apps & Marketplace and locate the Helm Chart for Hammerspace CSI, as shown in the image below.

 

Hammerspace CSI Driver

Hammerspace CSI driver supports block, file, and shared storage. By offering three different ways of provisioning storage virtually any use case or application is covered. There are some important distinctions between the three options. Block persistent volumes can be shared between pods, while a file-backed persistent volume provides a local non-sharable file system. A shared persistent volume is an NFS interface that allows pods to actively share file data, both locally and globally. Block and shared volumes can be shared across any number of pods across any distance on any infrastructure, whether on-premises or in cloud. The following image is a good illustration of the Hammerspace CSI driver architecture and the completeness of what it offers.

 

Hammerspace Installation with Helm

The Helm Chart also contains a README that lists prerequisites and detailed descriptions of the three different ways you can provision storage for a Rancher cluster.

 

Block Storageclass

The Helm Chart puts great powers in your hands to control exactly how storage is provisioned. You have full control over things such as export options, volume expansion, as well as declarative policy control over data placement, geolocation, protection, and virtually anything imaginable. You can, of course, accept the default settings to make is super simple. Block storage is ideal for SQL databases or other applications that require a simple block interface.

 

File-backed Storageclass

There are multiple potential use cases for File-Backed storage. NoSQL databases, such as MongoDB, that leverages its own replication mechanism for distributing data is a good example. MongoDB provides its own linear scalability which also limits the need for external scale-out storage.

 

NFS Storageclass

NFS shared storage is, of course, the ideal choice for Rancher Pods that need to share file storage without the pain, configuration complexity, and data-copy sprawl of replication. Why replicate data between volumes and layers when NFS solved that problem decades ago?

 

Summary

Rancher and Hammerspace, together, vastly simplify and accelerate the delivery of applications and data. The declarative nature of the solution is a truly powerful and innovative approach that makes applications and data always available, highly responsive, and globally portable. This is the optimal solution for CI/CD workflows.

 

Hammerspace reduces configuration complexity without losing the granular controls that administrators may want. Hammerspace Objectives are a simple and easy way to declare the intent of data or, in other words, tell the system what you want. You should expect nothing less of modern data services. This is distinctly different from aging storage paradigms that require administrators to tell a storage system not only what but also spell out in painstaking detail how it must be done. Declarative controls empower administrators to accomplish more with less. Smart and well-designed software knows how to accomplish a declared intent.

 

Johan Ballin

Director of Technical Marketing