Development environment into Docker

[Sonlis]

During the first coding evening, we realized that setting up the development environment was quite cumbersome and not that beginner-friendly. One idea that came to mind was to build a Docker image with the development environment ready. I am starting a discussion here about ideas that came to mind, please share your thoughts about them or bring up more solutions.

1. Application + database into one Dockerfile.

We could build an image based on node.js and install a mongo db database in it. This would makes the deployment on local environment the easiest possible: download the image, the app is running. Also, if the image is stored in a private repository, then environment variable can be automatically configured. I see 3 disadvantages though:

• Image size becomes bigger by having everything inside.
• The image would need to be built every time the dev branch is modified. I assume automatic CI would be a solution here, for example Github Action has a free tier up to 2000 minutes / month.
• Where to store the image. Either environment variables are straight up configured in the Docker, hence it needs to be in a private repository, or public repository, but people need to configure the variables themselves which is not a so friendly solution. Dockerhub offers 1 free private repo, so we could create a HackJunction user, store the images there and login to this user when we want to pull development image.

2. Application and database in different Dockerfiles, docker-compose to set up environment.

This would solve the problem of image size + would follow best practices to break down the application into smaller pieces. Create a docker for the application, and pull a MongoDB image. Running a docker-compose up would setup the environment and work out of the box.

Advantages

• Running docker-compose up and docker-compose down would make the environment spin-up and tear-down very easy.
• HackJunction docker image easier to maintain.

Drawbacks

• Same problem than with the previous solution regarding image storage and maintenance.
• Require to install Docker AND docker-compose.

3. Only use database container.

In this scenario, the code is still downloaded locally, but we provide an image for MongoDB, so we do not rely on external services. A developer could download the image, and by setting the right URL in .env files, have access to it. It can also be spin-up and tear-down with ease.

Advantages

• No need to maintain a private image for JunctionApp.

Drawbacks

• "Easy" solution and does not open to future possible project i.e containerization of the master app.
• Setup for junction app has to be ran locally, and some npm commands take a while to complete.

Please let me know what you think about this :)

[Sonlis]

Our biggest problem was to configure a database up and running that the application could connect to. So, I created a very simple docker-compose file to set up and tear down easily a database that the application can connect to.

Instructions.

You need Docker and Docker-compose. If you are on Mac or Windows, when installing Docker, you automatically get Docker-compose with it.
To install docker, follow this link and select your operating system. It will install docker.
If you are on a linux system, you will need these extra-steps: install docker-compose for linux.

Once this is done you will need a file called docker-compose.yaml in the base directory. Create the file, and copy paste the following content into it:

version: '3'
services:
    mongodb:
        image: mongo:3.6.23
        ports:
            - 27017:27017
        environment:
            MONGO_INITDB_ROOT_USERNAME: oui
            MONGO_INITDB_ROOT_PASSWORD: password

You can select any username and password you'd like.
Save the file. In backend/.env, modify the MONGODB_URI variable:

mongodb://oui:password@localhost:27017/admin

If you changed the username and password in the docker-compose.yaml file, match them accordingly in the env file.
Finally, when you need the database, just run from the base directory:

docker-compose up

Once you do not need the database anymore, run

docker-compose down