DockerEntrypoints

DockerEntrypoints refers to the ENTRYPOINT instruction in Dockerfiles, which configures the container’s main executable when it starts. It is designed to work with CMD, which provides default arguments, and both can be overridden at runtime with docker run options.

There are two forms of ENTRYPOINT. The exec form uses JSON arrays and runs the command directly, without a shell, for example: ENTRYPOINT ["executable","arg1","arg2"]. This form runs as PID 1 inside the container and handles signals properly, which is important for graceful shutdown. The shell form uses a single string, for example: ENTRYPOINT "executable arg1 arg2", and runs through /bin/sh -c. This form invokes a shell, enabling shell features but complicating signal handling and process management.

CMD acts as default arguments to the entrypoint when both are defined. If you provide additional arguments at docker run, those replace the CMD defaults. If no ENTRYPOINT is defined, CMD specifies the command to run. You can override the entrypoint at runtime with docker run --entrypoint, in which case the original CMD is treated as the arguments to the new entrypoint unless you supply new arguments as well.

For example, an image might use ENTRYPOINT ["python","server.py"] with CMD ["--port","8080"] to run a server by default, while allowing callers to override the port by supplying different arguments. A startup script as the entrypoint can initialize the environment and then exec "$@" to hand control to the final command, combining flexibility with proper signal handling.

Best practices favor the exec form for the main entrypoint, minimize shell usage, and prefer a small wrapper script that performs initialization and then executes the intended application.