Deploying
Reaction is an open source multi-service platform. All services are either publicly available Docker images on DockerHub or are Git repositories that include a Dockerfile so that you can build your own Docker image after customizing the service. This means that you can deploy the Reaction system on any infrastructure that supports running a Docker container cluster with a shared network.
Docker Images
You can find the published Reaction images on our DockerHub page.
Container Requirements
We recommend that you make at least 2GB of memory available to each container.
MongoDB Database
In a development environment, a local MongoDB database is created for you. In production, you will need to set up a separate production-ready MongoDB server and create a database in it. You then provide this database connection string in an environment variable for services that need to connect to it.
Because some services use change streams, your MongoDB server must be part of a replica set and must have oplog enabled.
Don't forget to set up periodic database backups. Do them as often as you can afford to do.
Postgres SQL Database
Reaction uses Hydra for OAuth2 login, and Hydra requires a Postgres SQL database. In a development environment, a local database is created for you. In production, you will need to set up a separate production-ready Postgres server and create a database in it. You then provide this database connection string in an environment variable for the hydra service.
Don't forget to set up periodic database backups. Do them as often as you can afford to do.
Database Migrations
When you deploy, the services will not start properly unless your Postgres and MongoDB databases are on the expected data version. Use the Reaction migration tools to migrate your data up or down before deploying. In some cases, migrating down is not possible. If you are rolling back to a previous Reaction version and migrating down is not possible, you will need to restore from the last backup prior to the upgrade.
Environment Variables
Most services need some additional information to run. In particular, they need to know what other services you are running on the network, and what the service names, ports, and other URLs and connection strings are. You provide this information by setting environment variables within each Docker container.
To determine what environment variables need to be set, look at the .env.example file in each project that you plan to deploy. Alternatively, you can start the container without any environment variables set, and then view the logs to see a list of environment variables that are required.