Docker is a PaaS solution. It hit the market in 2013 and has become the most popular solution nowadays for application containerization. It allows you to freeze the state of an application with a series of commands and recreate it on any other hardware that has Docker installed.
Docker has a simple declarative language
Most people use Docker to recreate a version of their software in the same state on any other computer. Docker images are created using Dockerfile files. These files use instructions such as:
ENV PYTHONUNBUFFERED 1
RUN apt-get update
# Pozwala dockerowi na zapamiętywanie zależności w naszej aplikacji
COPY ./requirements.txt requirements.txt
COPY ./requirements requirements
RUN pip install -r requirements.txt
# Dodaj naszą aplikację do kontenera
COPY . code
RUN python manage.py collectstatic --noinput
CMD ["sh", "-c", "gunicorn app.wsgi -b :8000 --log-file -"]
These are just some of them and the most popular. With them we can download another container created by another user using FROM, copy data using COPY from a local file source that is our own computer, execute a series of commands using RUN. Each command will create incrementally successive layers of the docker container. The most interesting thing is that when building the container again and changing the RUN command, the entire previous series of commands is remembered accordingly. Thanks to this, after minor changes, containers can be built really fast. At the very end, we can set the CMD that will be run when the container starts and this can be, for example, a script that starts a web server.
The Docker command line allows advanced operations on containers
Using the Docker interface, we can perform operations such as building a container or launching it. We can log in to other Docker providers and push the built container to them. We can version our containers so that we can revert to an older version of the container at a critical moment.
docker build -t [username/]<image-name>[:tag] <dockerfile-path>
docker push [registry/][username/]<image-name>[:tag]
The whole solution allows us to build software in a modern way and quickly exchange best patterns. Most open-source providers have ready optimized versions of their software in the form of Dockerfiles, so we can quickly check them out in an isolated and independent way from our software. It's important to remember that docker containers are independent of each other and have a file system separate from each other. If we do not allow it, they do not share any files with each other. This avoids version conflicts and allows us to focus on developing quality software.