This document provides basic information about how we leverage Docker on Shopsys Framework to make the work with it as easy as possible.
These days, developers use machines with different settings and configurations to work with. Modern platforms and frameworks have in most cases different software requirements, so their installation may take up to several hours. Our goal is to facilitate the installation of Shopsys Platform, ideally, if the developer does not need to significantly customize the configuration of his machine but instead of that, he has the possibility to install the platform immediately even with all dependencies - such as a virtual container. As a result, the developer can start development after a few minutes after downloading the platform.
2. Solution - Docker¶
Docker provides some resources to create isolated software containers based on the image, that includes everything needed to run it: code, runtime, system tools, system libraries, settings. Thanks to that, the containerized software will always run the same, regardless of the environment.
3. Shopsys-framework on Docker¶
Docker-compose.yml file contains the definitions of containers that will be created for the installation of the Shopsys Platform. Containers can communicate with each other. Data need to be saved into volumes, otherwise, these data will be lost after the container is closed.
Examples of containers that we use:
The definition of container consists of some options:
- image: Docker image that will be downloaded from Docker Hub and used
- build: data used for building the Docker image locally
- container_name: name of created container
- working_dir: file structure of created container
- volumes: location of data storage in which the data will remain even after the container is removed (see Volumes official docs.)
- ports: ports mapping, in default configuration is port 8000 mapped on port 8080 inside container
- environment: environment variables, after setting they can be used throughout the container
The definition of volumes, example:
- shopsys-framework-sync : name of the volume
- external: true : flag, which indicates, that volume is already defined external (in this case in docker-sync.yml)
Docker-sync.yml file contains the definition of synchronization for the docker-sync tool (it's used for Mac only).
Dockerfile is a text document that contains all the commands a user should call on the command line to assemble an image. It is like some kind of recipe by which final image is cooked.
Dockerfile example command:
FROM instruction specifies the base image, from which you are building
The official list of Dockerfile commands can be found on Dockerfile reference.
Php-ini-overrides.ini file contains php custom configuration which is used after installing php in a new container
Nginx.conf file contains Nginx configuration for new webserver container.
3.6 Images Distribution¶
While running Shopsys Platform on docker we needed to decide which distribution will be our images running on. We use 2 types of distributions for our images.
- Alpine Linux
Both have their advantages and drawbacks
Debian is one of the oldest distributions, so even users without Docker experience are able to modify these images because installing tools in this distributions is easy, common and very easy to find on internet if some problem occurs.
Also, Debian supports big amount of Databases and tools used by developers and are easy to install.
Debian image is pretty big, which is not ideal for pulling these images on Cloud Solutions because there is more resources used to pull these.
Alpine Linux is really small and minimalistic which means that clean installation of Alpine linux takes 6MB of memory which is great for pulling these images on Cloud where resource used for pull matters.
Dockerfiles based from Alpine are harder to modify because it is so slim that with every technology, database or tool you want to install, you also need to install all its dependencies, making it a bit hard to find right solutions.
Alpine Linux also has longtime problem with connection to MSSQL databases, due to missing support by Microsoft drivers.
We divide images used by us into 3 types.
Not extended, only used (PostgreSQL, Redis, Elasticsearch)¶
Those images are only used as it is, we are not extending them and we are only using them as they are. In this case we choose Alpine Distribution thanks to its size.
In this case, we use Debian mainly because we suppose users to modify these images often(adding php-extensions, implementing connections to Databases). As Debian is much more easier to modify, we decided that it will be best for new users to start on Debian, and if they care about size, they can always rewrite their Dockerfiles to be use alpine image if needed. Also many of clients using Shopsys Platform often connects to MSSQL databases and we want to make it easy for them.