docker run Command Reference

By Kaveh Alemi·Latchkey

Create and start a container from an image.

docker run is docker create plus docker start in one command. It pulls the image if missing, creates a writable layer, and runs the image command or one you pass. In CI you almost always want --rm so containers do not pile up.

Common flags

--rm - remove the container automatically when it exits
-e KEY=value - set an environment variable (repeatable)
-v src:dst - bind-mount a host path or named volume
-w, --workdir - working directory inside the container
--network - attach to a network, e.g. host or a named network
--entrypoint - override the image ENTRYPOINT
-d, --detach - run in the background and print the container ID

Example

shell

docker run --rm \
  -e CI=true \
  -e NPM_TOKEN=${NPM_TOKEN} \
  -v "${PWD}":/app \
  -w /app \
  --entrypoint sh \
  node:20 -c "npm ci && npm test"

In CI

Avoid -it in pipelines: the -t (TTY) flag causes "the input device is not a TTY" with no terminal attached. Always pass --rm so ephemeral test containers are cleaned up. Mount the workspace with -v and set -w to run tools against your checked-out code.

Key takeaways

Use --rm in CI so one-shot containers are removed on exit.
Drop the -t TTY flag in non-interactive pipelines to avoid "not a TTY" errors.
--entrypoint overrides the image default to run an arbitrary command.

Frequently asked questions

docker run Command Reference?

In CI?