In this lesson, you will learn how to use the VOLUME instruction in a Dockerfile to designate specific directories as persistence points, ensuring your critical data survives even when a container is removed.
The Problem: Ephemeral Layers
As we have learned in previous lessons, containers use a writable layer on top of read-only image layers. While this is great for lightweight, immutable deployments, it presents a challenge: when a container is deleted, everything inside that writable layer—including your database files, user uploads, or logs—is permanently destroyed. If you want your application to "remember" data across container restarts or removals, you must move that data outside of the container's lifecycle.
Defining Persistence with VOLUME
The VOLUME instruction tells Docker: "The files generated in this specific directory should be handled differently." When you define a VOLUME in your Dockerfile, Docker creates a mount point within the container. When the container runs, the data stored in that directory is bypassed by the Union File System's copy-on-write mechanism, allowing that data to persist independently of the container's writable layer.
Here is how you define it in your Dockerfile:
FROM alpine:latest
# Designate /data as a persistence point
VOLUME ["/data"]
CMD ["sh"]
Why use this in the Dockerfile?
By declaring /data as a VOLUME, you are providing metadata for any developer or orchestrator using your image. It acts as a clear signal that the application expects to write stateful information there. When you spin up this image, Docker will automatically manage a host-side directory for that path. Even if you delete the container, the data remains on the host, and you can map a new container to that same volume to restore the "memory" of your application.
This approach separates your application code (the immutable image) from your application state (the persistent volume), which is the cornerstone of building resilient, stateful services in any containerized environment.
Key Takeaway: The VOLUME instruction designates specific filesystem paths for data persistence, allowing critical information to survive the destruction of a container's writable layer.