Exec form vs. Shell form?

The ENTRYPOINT instruction in a Dockerfile can be specified in two forms: Exec form and Shell form. Here’s a comparison of the two:

Exec Form

• Syntax:

  ENTRYPOINT ["executable", "param1", "param2"]

• Behavior:
  • The command is executed directly without a shell.
  • It does not invoke a shell, which means that shell features (like pipes, redirection, etc.) are not available.
  • It allows for better signal handling, as the executable receives signals directly.
• Use Case:
  • Recommended for most cases, especially when you want to ensure that the command runs as intended without shell interpretation.

Shell Form

• Syntax:

  ENTRYPOINT command param1 param2

• Behavior:
  • The command is run in a shell (/bin/sh -c).
  • This allows the use of shell features like pipes, redirection, and environment variable expansion.
  • However, it can lead to issues with signal handling, as the shell process receives the signals instead of the executable.
• Use Case:
  • Useful when you need shell features or when you want to run multiple commands in a single line.

Summary

• Exec Form: Direct execution, better signal handling, no shell features.
• Shell Form: Runs in a shell, allows shell features, but may have signal handling issues.

In general, the Exec form is preferred for defining ENTRYPOINT in Dockerfiles.