What’s your preferred workflow for setting up Docker in development and production?

When setting up a new project, I typically create a designated folder, craft my Dockerfile, and set up the necessary files. In my docker-compose setup, I mount my local app folder and run it in development mode, which allows for automatic reloading on file changes. I often use SQLite for initial development since it's lightweight, but I implement an ORM to facilitate easy transitions to PostgreSQL in production. During testing, I usually wipe the DB file and restart the container for a fresh instance, all managed through a single compose file.

For development, I keep things pretty straightforward—just run the app directly with a local database, similar to what you're doing. However, Docker shines in production settings. I prefer using a single docker-compose.yml file with profiles for development and production instead of creating separate files; it keeps everything clean and organized. The trick is in configuring your volume mounts correctly to avoid issues with node_modules being overwritten. Also, I highly recommend utilizing multi-stage builds to maintain slim production images.

The node_modules sync issue is a common headache! You can resolve it by using a named volume configuration, like this:

volumes:
- .:/app
- /app/node_modules

This keeps your dependencies within the container separate from the host's bind mount, helping to avoid those sync issues. For separating development and production setups, I use a base compose.yml file with overrides rather than having multiple complete files, which helps with cleanliness.

I prefer to maintain two separate compose files for development and production, and often a third for staging deployments. Each environment has its dedicated database container, and I specify which one to connect to right in the compose configurations. My development happens on my laptop, staging on a server, and production can either be cloud-based or on a server, depending on the project. I usually push updates to staging first, even without testing there manually, and then merge into the main branch. Before starting a major task, I clone the production database to refresh my local development one to ensure everything aligns perfectly. I've also built a local web UI tool to back up all my PostgreSQL containers!