Container Native Storage - Solving the Persistent Storage Challenge with Gluster

Containers are designed to be ephemeral, stateless entities, but distributed applications running inside container clusters need a place to persistently store their data. Container orchestration systems like Kubernetes / OpenShift do meanwhile offer APIs for provisioning and providing persistent storage, but it is quite cumbersome to use them manually. This presentation will introduce the problems with persistent storage for containers, will describe Kubernetes storage mechanisms, and will then explain how Red Hat has solved the persistent storage challenge by bringing Gluster into the container cluster: Gluster is a scale out, distributed software-defined storage system, that can now run in containers. In the Container-Native Storage (CNS) solution, Gluster runs hyper-converged in a Kubernetes cluster and dynamically provisions persistent storage volumes of desired size upon user request. The presentation will include demos of how to stand up a CNS setup with a single command and how to dynamically provision storage for containerized applications. The session will explain the inner workings of this completely open-source software solution and conclude with a roadmap.