Skip to content

Latest commit

 

History

History
197 lines (167 loc) · 7.46 KB

README.md

File metadata and controls

197 lines (167 loc) · 7.46 KB
Raw

Deploy docker containers

StoryWeaver has five services: The storyweaver application (Rails backend + React frontend), Postgress database, couchbase, elasticsearch, and nginx. This documentation will lay out the steps to deploy StoryWeaver on your server.

Prerequisites

  • Create required number of vps running ubuntu 18.04. If you are going to create one for each service go with five or if you are going for single system install, create one. Note that elasticseach requires atleast 4GB of memory so please keep in mind when creating servers.

  • Open ports TCP 2377, TCP and UDP 7946 and UDP 4789 on all machines. These ports are used by docker to communicate between nodes. Make sure that the port '80' and '443'(for https) is open and accepting connections on the machine intended to run 'nginx'.

    On aws, opening ports can be managed by editing the instance's security group. See documentation here.

    On google cloud platform, the necessary firewall rules need to be created and assigned to the instances. See here.

    Documentation for digital ocean can be found here

  • SSH_KEY : An ssh key that can be used to access the machines created above. You can create a new key or use an existing one. The steps for creating a new key or using an existing one differ from provider to provider. Documentation for AWS and GCP can be found below.

Machine Setup

  • Install git and docker on all machines. docker version needs to be 18.09.1-rc1 or greater.

    docker:

    If docker is not already installed then:

      curl -fsSL https://get.docker.com -o get-docker.sh
  CHANNEL="test" sh get-docker.sh

    This will also install git. If you are not using this method, you can install git using

     sudo apt-get update 
 sudo apt-get install git

    If docker is already installed and is running a lower version, then do the following to update

      # Uninstall current version
  sudo apt-get purge docker*

  # Install latest version
  curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
  sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu bionic test"
  sudo apt update
  sudo apt install docker-ce

    Make sure of the install by running

      docker --version
  git --version

  • Place the ssh key to server supposed to run nginx service

Deployment

This provides dockerized environment for developing 'sw-core' and 'sw-web' projects.

  • ssh into the machine where rails should be run. Let's call this the 'rails server'

  • Clone application repositories to a directory called 'workdir' in rails server (This can be any directory. see "WORKDIR" in config section given below)

    mkdir workdir
cd workdir
git clone https://github.com/PrathamBooks/sw-core.git
git clone https://github.com/PrathamBooks/sw-web.git
# Make sure "gem tzinfo-data" in Gem file

  • Edit sw-core/config/database.yml add a new line under 'default' "host: sw-postgres" in it.

    #default looks like after edit
default: &default
adapter: postgresql
encoding: unicode
pool: 5
database: spp
timeout: 5000
username: spp_user
host: sw-postgres

development:
<<: *default
...
...

  • ssh into the machine where nginx should be run. This node will be used as the docker swarm manager. Let's call this the 'nginx server'

  • Clone the deployment repository

      git clone https://github.com/PrathamBooks/sw-docker.git
  cd sw-docker

  • Edit the 'dev-config' file and set appropriate values. - Edit the 'config' file and set appropriate values.

    • 'USER': Set this to the username which will be used to ssh into the machines (The default user for all ubuntu machines in aws is 'ubuntu').

    • 'SSH_KEY': The path to the ssh key placed on the nginx server

      for example: If the private key file id_rsa is placed at /home/ubuntu/ in nginx server, then the value should be SSH_KEY:/home/ubuntu/id_rsa

    • 'SHARED_DIR': Set this to a path on the machine going to run rails application. This path does not have to exist as it wil be automatically created on that machine. This is used by the rails application to store logs and other data. For example, setting it to below value will create a directory named logs in the machine running rails and mount it inside the container. The application stores logs inside this directory, which can then be accessed directly from the host machine without going inside the container.

        SHARED_DIR=/home/ubuntu/logs

    • The names of the services should be set to the IP addresses of the instances which should run them. An example config setup will look like

      ELASTIC=192.168.1.110
SW_APP=192.168.1.121
NGINX=192.168.1.215
POSTGRES=192.168.1.12
COUCH=192.168.1.3

      You don't necessarily need to use individual machines for all services. The multiple machine setup allows the flexibility of using the same machine to run multiple services. To do this set the config keys of these services to the same value. For example, for a three machine setup which runs elasticsearch on its own machine, nginx and rails both on another machine and postgres and couchbase together on another, then their values in the config would look like

        ELASTIC=192.168.1.1
  SW_APP=192.168.1.2
  NGINX=192.168.1.2
  POSTGRES=192.168.1.3
  COUCH=192.168.1.3

    • The other keys in the config are for using third party

    Difference: 'WORKDIR' is the absolute path to the directory where 'sw-core' and 'sw-web' are cloned. This directory will be mounted into the rails container, so, any files put inside this directory will be available inside the container.

  • Run

     bash dev-deploy.sh

  • Login into rails container and build react app

      docker exec -it <rails containerID> bash
  bash /workdir/sw-core/lib/scripts/build_script.sh
  #update localhost:3000 to <nginx ip address> in build script

    The rails server will take some time to come up. You can check the progress by running the following.

      docker logs <rails container name>

    Wait for sometime and then go to the ip address of nginx server

Known issue

In very rare cases just after deployment, containers can have problems connecting to each other due to internal DNS resolution issues with docker swarm. To mitigate this, we check for the issue after deployment and if it has happened, the containers are redeployed.