Having a clean and productive workflow is essential for every developer, Docker and docker-compose are just the right tools for keeping it clean and avoid installing anything locally.
For the last 15 years I had to work on multiple projects for different clients at the same time, having to install locally different Mysql versions, PHP versions, Java, Nodejs, you name it, obviously this became a real nightmare as the projects kept rolling in. At one point I tried to create a VM for each project using the old way, but I ended up consuming my HD space really fast so it wasn't a long lasting solution.
Adding to the mess of different versions of a particular db, dev lang, etc. If I need to setup an environment for a project, let's say Mongo, Neo4j, Go, Reactjs, I would have to install everything locally, the nightmare continued.
Here is where Docker and docker-compose comes to save the day.
Docker is a great tool for working with containers, isolating the app, db, etc. It enables painless installation of services and sharing development environments with others without the hassle of getting the same version of everything and installing the tools locally.
Lets see with a small example how I work with these two amazing tools to setup a working environment that I can easily share with other developers.
In this example will work with Neo4j and Go.
Create a project folder, lets call it docker-go-neo4j, in the root of that directory create a docker-compose.yml file with the following
version: "3.2" services: neo4j: image: neo4j:latest container_name: 'neo4j' volumes: - '$HOME/neo4j/data:/data' - '$HOME/neo4j/conf/:/conf/' ports: - 7474:7474 - 7687:7687 restart: on-failure networks: - neo4j_go_net command: neo4j networks: neo4j_go_net: driver: bridge volumes: dgraph:
I like to create seperate networks for each environment that I setup, now save the file and run the following command from the root of your projects folder.
docker-compose up -d
This will download the image for Neo4j and starts the container in the background (remember this command, we will run it each time we add a new service to our docker-compose.yml file
Verify if the containers are running
You should see neo4j listed
TIP: You can verify that the database is working by heading to http://localhost:7474
Lets add another service inside the docker-compose.yml, this service will depend on the neo4j service we just created. Add the following bellow the neo4j service
backend: container_name: 'api-go' build: './backend' ports: - '8080:8080' volumes: - './backend:/go/src/app' depends_on: - 'neo4j' networks: - neo4j_go_net
The backend service will use a Dockerfile to setup go and copy the local files into the container, install the dependencies and setup hot reloading with the fresh package, so before running any other command, create a backend folder and inside of that folder create a file named Dockerfile without any extension.
FROM golang:latest WORKDIR /go/src/app COPY . . RUN go get github.com/pilu/fresh RUN go get -d -v ./... RUN go install -v ./... CMD [ "fresh" ]
what this file basically does is fetch the golang image from docker hub, setup a working directory, copy everything inside the backend folder into the working directory and install the needed packages, finally runs the fresh command for hot reloading.
Inside the backend directory you also need to create a main.go file and a public directory with an index.html file inside.
The complete source code can be viewed here docker-go-neo4j
The main.go and index.html files are log so I suggest you check them out in the repo, we are going to be working with the sample code from Neo4j, this way we know everything works.
The Idea behind this was not to create and application from scratch but for just illustrate how you can setup a development environment with Docker and docker-compose to work with Neo4j and Go.
Let's wrap it up
Hope this gives you an idea on how to setup your working enviroments and how easy is to share this with other so they can have the same environment on their machines. If you're new to Docker and the world of containers, give it a try, if not for production at least for development, this will change your life and will make your workflow much faster, avoid the nightmare of installing everything locally.
Checkout the complete sourcode on Github