Generally created with the filename “Dockerfile” and saved in the root of the project
Another name can be used as well but during build the file name needs to be specified explicitly using the -f flag
Every Docker command that we write is going to be created as a new layer for the image
When using FROM for getting a base image whatever is specified in the base image is inherited into our target image as well
Commands
FROM <distro-name>The FROM command is an required command in every Docker file
It is the base image that is going to be used for making the new image
If we don’t want to start from on image then “FROM scratch” can be used
WORKDIR <directory-path>Similar to running cd <path>
Recommended way of change the working directory when building Docker Images
ENV <key>=<value>Environment Variables can be optionally set which can be used later on in the docker file They can be accessed and modified in the container after the image is built
ENV can be created without the = signal as well
ARG <key>=<value>Arguments are a special type of environment variable that is only available during the build process of the image.
To pass values for the arguments during build use the --build-arg flag. If a FROM command is encountered in the Dockerfile all the arguments that have been defined before it will be reset
Additionally ARGs can be declared globally before the 1st FROM statement but to use these values the variables need to be called in each of the stages as required
ARG FILENAME=foo.txt
FROM ubuntu:latest
ARG FILENAME
RUN echo "Hello World" >> ${FILENAME}ARG can be used to dynamically select the version of the image to be used
COPY <host-system> <docker-image>Copy files/ source code from host system into the Docker Image
RUN <command>Used to run some shell commands that can be used to install/ unzip/ update various binaries that are present in the image at this point of time
Multiple commands can be changed together using the && operator
EXPOSE <port-no>Used to specify the ports that need to be exposed by this container so the user can connect to them using the ‘-p’ flag
By default no ports are exposed on a container
CMD ['command', 'to', 'run']It is a required parameter and appears at the end of the docker file
Specifies how the container needs to be run
Only one CMD is allowed per docker file so if multiple are specified the last one wins
USER <username>Run commands in the image as a specific user
Some images come with an non root user created out of the box
When using COPY and WORKDIR commands add the --chown flag to ensure the operations are being performed as the non-root user
ENTRYPOINT ['entry-point.sh']The actual syntax of the command run at the start of the container is ENTRYPOINT CMD
It can be used to specify an script that needs to be executed every time at the start of the container