Skip to main content
In this lesson we’ll cover the storage fundamentals you need to run stateful workloads with containers and Kubernetes. Understanding Docker storage concepts first makes it much easier to follow how Kubernetes extends those ideas with the Container Storage Interface (CSI), PersistentVolumes (PV), and PersistentVolumeClaims (PVC). Two core concepts define storage for containers:
  • Storage drivers — implement how a container’s writable layer and image layers are managed on the host filesystem (examples: overlay2, aufs, devicemapper). Storage drivers operate at the container runtime level and determine how container filesystem changes are isolated and persisted on the host.
  • Volume driver plugins — provide integration with external or specialized storage backends (examples: NFS, cloud block storage, third-party plugins). Volume drivers enable containers to use persistent storage that outlives the container process and can be shared, mounted, or managed by external systems.
We focus on Docker storage drivers here as a foundation. If you already completed the Docker Training Course for the Absolute Beginner, you may want to skip this section; otherwise, this will help you later when learning Kubernetes storage patterns such as CSI, PV, and PVC.
A dark presentation slide titled "Docker Storage" showing two rounded boxes labeled "Storage Drivers" (left) and "Volume Drivers" (right).
If you already know Docker’s storage drivers and volume plugins, you can skip this lesson. Otherwise, continue — understanding these concepts will make Kubernetes storage much clearer.

How storage drivers and volume plugins differ

  • Storage drivers (runtime-level) manage the container image layers and the container’s writable layer. They determine copy-on-write behavior, layer merging, and how changed files are stored on the host.
  • Volume plugins (external backends) expose persistent mounts to containers and let you attach storage that survives container restarts, can be shared between containers, or implemented by remote systems.
Key examples:
  • Storage drivers: overlay2, aufs, devicemapper
  • Volume plugins/backends: NFS, Amazon EBS, Google Persistent Disk, GlusterFS

Quick comparison

Inspecting Docker storage on a host

To see which storage driver Docker is using:
  • Run:
  • Or to print only the storage driver:
(Use the above command to parse the Driver field programmatically.)

Why this matters for Kubernetes

Kubernetes separates concerns further:
  • Container runtimes still use storage drivers to manage image and writable layers on nodes.
  • Kubernetes uses PersistentVolumes (PV) and PersistentVolumeClaims (PVC) to provide portable, declarative, and managed storage for workloads.
  • The Container Storage Interface (CSI) standardizes plugin development so external storage can be used consistently across container orchestrators.
If you understand how Docker storage drivers isolate and persist changes, it becomes straightforward to map those behaviors to Kubernetes PV/PVC semantics and CSI-based storage drivers.

Further reading and references

Next, we’ll dive into how Kubernetes uses these abstractions and how CSI plugins connect cluster workloads to persistent storage backends.

Watch Video