Backend of a simple, quick and scalable RESTful Socket.io/redis/express application. Originally made for Reaktor's summer internship challenge. They ended up hiring me, in case you're curious. :)

See frontend here: https://github.com/melonmanchan/reaktor-chat-client

The server has a single master node and X workers working in a clustered manners. The amount of workers is configured by the environment variable WORKERS. By default, it is the amount of processors on the machine minus one. All outfacing HTTP requests and socket connections are then served by these workers. Sticky sessions are implemented (request from same IP address) will end up into the same worker. This is required by Socket.io's HTTP fallback mechanism. If using strictly websockets, just Redis would be enough.

Each piece of the cluster has an in-memory data-store (see app/store/chatstore.js) The application state such as the connected users and what channels they're currently part of is held there. The master node of the cluster is the 'Single Source of Truth', which has the final say in what the application state looks like. Upon startup, a worker node requests the current state from the master node. In this way, even if a worker crashes and has to be restarted, it's state will still stay in sync with the rest of the cluster. Each store also holds the sockets connected to that particular node, mapped to the users. Whenever a state-manipulating event happens (user logs in, joins a channel etc.), the worker the user socket is connected to fires an event, which causes the rest of the cluster to update their state.

Further reading on clustering socket.io, which this application attempts to expand upon: https://github.com/elad/node-cluster-socket.io

First off, install Yarn: https://yarnpkg.com/ Then, clone this repository and install the required packages

    git clone https://github.com/melonmanchan/reaktor-chat-server
    cd reaktor-chat-server
    yarn

Then, you should install Redis. Redis is used to store chat messages and user credentials, and exchange messages between workers and clusters.

https://redis.io/download

Afterwards, launch Redis on localhost:6379 and run npm start to start the server. See the configuration section below for various tuning options. Just one worker cluster is more than enough for running in development

    WORKERS=1 npm start

Docker, orchestrated by Docker Compose, is the recommended way of running this application in production. Simply install those two:

https://docs.docker.com/engine/installation/

https://docs.docker.com/compose/

and change some of the default environment variables, such as the JWT token, in the file chat.env. Afterwards, run:

    docker-compose up

In the root of this repository. This will pull the required images from docker hub and start the application on port 8000.