Dockerfile Layers Explanation

Dockerfile 
FROM ubuntu:24.04
RUN apt update
RUN apt install -y golang
WORKDIR /here
COPY ./server.go ./server.go
CMD ["go", "run", "server.go"]

alt text

Layer-by-Layer Explanation

Layer Dockerfile Instruction What happens internally Notes
Layer 0 FROM ubuntu:24.04 - Check Docker cache:
- If image exists locally → load it.
- If not → pull from Docker Hub.
- Then loaded Base image.		 Base layer for all subsequent layers.
Layer 1 RUN apt update - Create temporary container from Layer 0.
- Execute apt update.
- Commit container as intermediate image.
- Remove temporary container.		 Layer caches the updated package list.
Layer 2 RUN apt install -y golang - Create temporary container from Layer 1.
- Execute apt install -y golang.
- Commit container as intermediate image.
- Remove temporary container.		 Golang is installed and saved in the layer.
Layer 3 WORKDIR /here - Set working directory for subsequent instructions.
- Internally creates metadata for image.		 Does not create a real filesystem change.
Layer 4 COPY ./server.go ./server.go - Copy server.go from build context into /here/server.go in the container. Creates a new layer with the source file.
Layer 5 CMD ["go","run","server.go"] - Set default command for the container at runtime. Does not create a new image layer in most cases, just metadata.

intermediate layers build on top of each other until the final image.

Why Docker keeps intermediate images?

  • Every command that changes the filesystem (RUN, COPY, ADD) creates a new intermediate image.

  • These intermediate images are cached for faster rebuilds:

    • If you rebuild the image and the previous layers haven’t changed, Docker will reuse cached layers.
    • This makes builds faster because Docker doesn’t rerun commands unnecessarily.

  • By default:

    • If build fails, intermediate images are kept for debugging.
    • If build succeeds, Docker may automatically prune intermediate images unless cache is needed.