Ruby was one of the first languages to have support from MongoDB with an official driver. The official mongo-ruby-driver on GitHub is the recommended way to connect to a MongoDB instance.

Installation is as simple as adding it to the Gemfile:

gem 'mongo', '~> 3.4'

And then in our class we can connect to a database:

require 'mongo'
client = Mongo::Client.new([ '127.0.0.1:27017' ], database: 'test')

This is the simplest example possible, connecting to a single database instance called test in our localhost. In most use cases we would at least have a replica set to connect to, as in the following snippet:

client_host = ['server1_hostname:server1_ip, server2_hostname:server2_ip']
 client_options = {
  database: 'YOUR_DATABASE_NAME',
  replica_set: 'REPLICA_SET_NAME',
  user: 'YOUR_USERNAME',
  password: 'YOUR_PASSWORD'
 }
client = Mongo::Client.new(client_host, client_options)

The client_host servers are seeding the client driver with servers to attempt to connect to. Once connected, the driver will determine the server that it has to connect to according to the primary/secondary read or write configuration.

The replica_set attribute needs to match the replica set name to be able to connect.

user and password are optional but highly recommended in any MongoDB instance. It's a good practice to enable authentication by default in the mongod.conf file and we will go over this in Chapter 7. Monitoring, Backup, and Security.

Connecting to a sharded cluster is similar to a replica set with the only difference being that, instead of supplying the server host/port, we need to connect to the mongo router, the mongos process.