Working directory

Learn to set the working directory with -w.

Working directory

-w (--workdir) sets the starting directory for the process inside the container; it overrides WORKDIR from the Dockerfile. It is widely used to mount a folder and run a command there:

docker run --rm -w /tmp alpine pwd
# /tmp
docker run --rm -v "$PWD":/src -w /src node:20 npm install

The second command runs npm install in the project folder on the host without installing Node.

Overriding image settings

docker run can override most of the settings in the image: the command (arguments after the image), the working directory (-w), the entry point (--entrypoint), the user (-u) and environment variables (-e). Replacing the entrypoint with a shell to look inside an image is common:

docker run --rm -it --entrypoint sh api:1.0

Exercise Run a tool in the project

Write the command for a container, deleted when done, that mounts the current folder ("$PWD") at /src in the container, sets the working directory to /src and runs npm test in the node:20 image.

Test yourself

With WORKDIR /app in the Dockerfile, what does docker run -w /data img pwd print?
- /
- It errors
- /data
- /app
Show answer

/data — Run-time options override image settings.

Replace the image's entry point with a shell.

docker run --rm -it ____ sh api:1.0

Show answer

docker run --rm -it --entrypoint sh api:1.0

Match.

-w

Working directory
--entrypoint

Entry point
-u

User
arguments after the image

Command

Next lesson: Switching users

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies