Start a Local Instance of Postgresql

For a majority of applications you should just use Postgresql. I know I'm writing for an audience that loves their Datomics and XTDBs, but the rest of this is focused on the Postgresql life path.

If you get to the end I'm sure it would be clear how you would slot in your alternative database of choice.

• Make a docker-compose.yaml file and put the following into it.

  This will let us start an instance of Postgres on the correct version for everyone working on the code. In addition it can be helpful when there are changes that require a wipe or a restart to have all your data be in a project-local folder like ./data/db.

  services:
    postgres:
      image: postgres:17
      restart: unless-stopped
      env_file: ".env"
      healthcheck:
        test: ['CMD-SHELL', "sh -c 'pg_isready -U postgres -d postgres'"]
        interval: 3s
        timeout: 3s
        retries: 10
      environment:
        POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
      volumes:
        - ./data/db:/var/lib/postgresql/data
      ports:
        - '5432:5432'

• Make a file named .env and put the following into it.

  POSTGRES_PASSWORD=postgres

• Add .env to a .gitignore file.

  The rest of will try to not be git specific, but this one in particular is worth calling out. Accidentally commiting a .env file is one of those things that's easy to do by mistake if you are not careful. And its important not to, because you can leak API keys and similar that way.

  Also, while we're at it, add .cpcache, .clj-kondo/.cache, and data to your .gitignore. .cpcache folder comes from the clojure CLI tools and isn't usefully sharable between machines. .clj-kondo/.cache is similar temporary files from the linter. data is where all the data from the local postgres instance will go.

  .env
  .cpcache
  .clj-kondo/.cache
  data

• Run docker-compose up -d

  This should download start up postgres in the background.