In order to power modern microservices and other stateless applications, Kubernetes operators need to have a way to manage stateful data storage on the cluster. While it's advantageous to maintain as much state as possible outside of the cluster in dedicated database clusters as a part of cloud-native service offerings, there's often a need to keep a statement of record or state cluster for stateless and ephemeral services. We'll explore what's considered a more difficult problem in the container orchestration and scheduling world: managing locality-specific, mutable data in a world that relies on declarative state, decoupling physical devices from logical objects, and immutable approaches to system updates. We'll explore strategies for setting up reliable, replicated storage for modern database engines.

In this chapter, we will discuss how to attach persistent volumes and create storage for stateful applications and data. We will walk through storage concerns and how we can persist data across pods and the container life cycle. We will explore the PersistentVolumes types, as well as PersistentVolumeClaim. Finally, we will take a look at StatefulSets and how to use dynamic volume provisioning.

The following topics will be covered in the chapter:

• Persistent storage
• PersistentVolumes
• PersistentVolumeClaim
• Storage Classes
• Dynamic volume provisioning
• StatefulSets