Deploy to Fly.io with SQLite

note

To follow this how-to guide, you'll first need to install the Fly CLI and create an account by following this official guide. You will also need an existing Platformatic DB project, please check out our getting started guide if needed.

note

Navigate to your Platformatic DB project in the terminal on your local machine. Run fly launch and follow the prompts. When it asks if you want to deploy now, say "no" as there are a few things that you'll need to configure first.

You can also create the fly application with one line. This will create your application in London (lhr):

fly launch --no-deploy --generate-name --region lhr --org personal --path .

The fly CLI should have created a fly.toml file in your project directory.

Explicit builder

The fly.toml file may be missing an explicit builder setting. To have consistent builds, it is best to add a build section:

[build]
  builder = "heroku/buildpacks:20"

Database storage

Create a volume for database storage, naming it data:

fly volumes create data

This will create storage in the same region as the application. The volume defaults to 3GB size, use -s to change the size. For example, -s 10 is 10GB.

Add a mounts section in fly.toml:

[mounts]
  source = "data"
  destination = "/app/.platformatic/data"

Create a directory in your project where your SQLite database will be created:

mkdir -p .platformatic/data

touch .platformatic/data/.gitkeep

The .gitkeep file ensures that this directory will always be created when your application is deployed.

You should also ensure that your SQLite database is ignored by Git. This helps avoid inconsistencies when your application is deployed:

echo "*.db" >> .gitignore

The command above assumes that your SQLite database file ends with the extension .db — if the extension is different then you must change the command to match.

Change the connection string to an environment variable and make sure that migrations are autoApplying (for platformatic@^0.4.0) in platformatic.db.json:

{
  "db": {
    "connectionString": "{DATABASE_URL}"
  },
  "migrations": {
    "dir": "./migrations",
    "autoApply": true
  }
}

Configure server

Make sure that your platformatic.db.json uses environment variables for the server section:

{
  "server": {
    "logger": {
      "level": "{PLT_SERVER_LOGGER_LEVEL}"
    },
    "hostname": "{PLT_SERVER_HOSTNAME}",
    "port": "{PORT}"
  }
}

Configure environment

Start with your local environment, create a .env file and put the following:

PORT=3042
PLT_SERVER_HOSTNAME=127.0.0.1
PLT_SERVER_LOGGER_LEVEL=debug
DATABASE_URL=sqlite://.platformatic/data/movie-quotes.db

Avoid accidental leaks by ignoring your .env file:

echo ".env" >> .gitignore

This same configuration needs to added to fly.toml:

[env]
  PORT = 8080
  PLT_SERVER_HOSTNAME = "0.0.0.0"
  PLT_SERVER_LOGGER_LEVEL = "info"
  DATABASE_URL = "sqlite:///app/.platformatic/data/movie-quotes.db"

Deploy application

A valid package.json will be needed so if you do not have one, generate one by running npm init.

In your package.json, make sure there is a start script to run your application:

{
  "scripts": {
    "start": "platformatic db"
  }
}

Before deploying, make sure a .dockerignore file is created:

cp .gitignore .dockerignore

Finally, deploy the application to Fly by running:

fly deploy