Docerazing Rails with Nginx and Postgresql

Description

This repository is a playground for docerazing Rails, Postgresql, Nginx and Rspec with docker compose. The aim is creating a simple and robust development environment with docker compose and deploy the code to production with the help of docker swarm. This rails app is a simple poll application. It is listing polls and lets the user to vote for a poll. A user would vote only once with the same IP.

Development

Run $ docker-compose up -d and wait for all services are run and functioning. In order to use the app the db schema must be updated. $ docker-compose exec rails rake db:setup command will update the database and loads seed data. Now go to http://localhost:3000 and vote for some polls.

Deployment with VirtualBox as single node

1. $ docker-machine create -d virtualbox node1
2. $ eval $(docker-machine env node1)
3. $ docker swarm init --advertise-addr $(docker-machine ip node1)
4. $ docker stack deploy -c docker-stack.yml rails

Wait until all services up. You could check status of services with $ docker service ls command. This would take some time so be patient.

Seed the database with $ docker exec $(docker ps -lq --filter name=rails_rails) rake db:seed.

Do not forget that we are using newly created image as docker host. To use your host's docker as docker host just type $ eval $(docker-machine env -u).

To view rails app in action we need to know our server's ip. $ docker-machine ip node1 command gives us the public ip.

Deployment with VirtualBox as Swarm Cluster

We will create 3 servers such as 1 manager and 2 worker nodes through VirtualBox. So let's create servers first.

1. $ docker-machine create -d virtualbox node-1
2. $ docker-machine create -d virtualbox node-2
3. $ docker-machine create -d virtualbox node-3

We will use node-1 as manager node. So type as the following:

1. $ eval $(docker-machine env node-1) We change the host as node-1
2. $ docker swarm init --advertise-addr $(docker-machine ip node-1)
3. $ TOKEN=$(docker swarm join-token -q worker)

Our manager node is ready. Now the next is to join worker nodes to manager node. So type as the following:

1. $ eval $(docker-machine env node-2)
2. $ docker swarm join --token $TOKEN $(docker-machine ip node-1):2377
3. $ eval $(docker-machine env node-3)
4. $ docker swarm join --token $TOKEN $(docker-machine ip node-1):2377
5. $ eval $(docker-machine env node-1)

Our swarm cluster is ready. $ docker node ls command shows likely the following output.

ID	HOSTNAME	STATUS	AVAILABILITY	MANAGER STATUS	ENGINE VERSION
dwirgheqrjnz *	node-1	Ready	Active	Leader	18.09.6
qnmsyvcjf7r56	node-2	Ready	Active		18.09.6
w1vljpmj0rvcv	node-3	Ready	Active		18.09.6

We can deploy our docker-stack.yml file with $ docker stack deploy -c docker-stack.yml rails command. This will take a considerable time so use $ docker service ls command constantly so as to see the status of services.

We are almost finish. Once all services up and running we need to seed the database with the following command:

• $ docker exec $(docker ps -lq --filter name=rails_rails) rake db:seed

Last, to see the cluster in action get the node-1 machine ip with $ docker-machine ip node-1 and visit the poll application.

Deployment To AWS EC2 As Swarm Cluster

Deploying to AWS EC2 instances is a bit complicated. First of all, a custom VPC with its subnets and routing tables must be created. Here is a working configuration. I tested this configuration with eu-west-2 region and it worked.

Once the configuration is set up, the following command should create the manager instance:

$ docker-machine create -d amazonec2 \
--amazonec2-vpc-id <your vpc id> \
--amazonec2-subnet-id <your subnet id> \
--amazonec2-region <your region> \
--amazonec2-zone <your region zone> \
--amazonec2-security-group <your security group name> \
aws-manager-1

If this works in your scenario, you would apply to the same command in order to create worker instances.

Once manager and worker instances are created, ubuntu user should be added to docker group for every instances such as the following command:

$ docker-machine ssh aws-manager "sudo usermod -a -G dokcer $USER"

We can init swarm mode on our manager instance with $ docker swarm init --advertise-addr eth0. The rest of other things like joining workers is the same as the preceding virtualbox example.