What is the difference between the CMD and ENTRYPOINT instructions in a Dockerfile?

CMD (Command) instruction:

• Definition: The CMD instruction provides a default command for the container. This command executes when the container starts, but if another command is specified during container startup, the CMD command is overridden.
• Use case: Use CMD for applications that typically run with the same command, such as starting a web server.
• Example:

  Dockerfile
  FROM ubuntu
  CMD ["echo", "Hello world"]

ENTRYPOINT instruction:

• Definition: The ENTRYPOINT instruction also provides a default command for the container, but it is designed to make the container behave like an executable program. Unlike CMD, ENTRYPOINT is less likely to be overridden at runtime unless explicitly using docker run --entrypoint.
• Use case: Use ENTRYPOINT for containers designed to run a specific program where end users should not easily change the program, such as for database services or application servers.
• Example:

  Dockerfile
  FROM ubuntu
  ENTRYPOINT ["top", "-b"]

Combining CMD and ENTRYPOINT:

• Purpose: Think of ENTRYPOINT as the container's executable program, while CMD provides default arguments. If no command-line arguments are provided, the CMD arguments are used.
• Example:

  Dockerfile
  FROM ubuntu
  ENTRYPOINT ["sleep"]
  CMD ["5"]

  Here, if the user runs the container without parameters (e.g., docker run myimage), it defaults to sleep 5. If parameters are provided (e.g., docker run myimage 10), the 10 overrides the CMD-specified 5, running sleep 10.

With this approach, you can flexibly control the container's behavior while still offering users customization options without compromising the container's intended purpose.