SoftwareEngineering

Installation process

• Install vagrant and ruby
  • Vagrant is a virtual machine manager and ruby is the language/interpreter
• Install librarian-chef with gem install librarian-chef
  • librarian-chef manages chef cookbooks. These cookbooks are used to install dependencies on the VM.
• Download the required chef cookbooks: librarian-chef install
  • This downloads the cookbooks.
• Start virtual machine with vagrant up (Note, first time will take a while)
  • This will initially download the base box and run the provisioning scripts. This installs things from the cookbooks as well as whatever else is listed in the Vagrantfile.
• SSH into the virtual machine with vagrant ssh and run gem install bundler; rbenv rehash; bundle install from the rails project directory.
  • Ruby libaries/dependencies are managed using RubyGems via the gem command. bundler installs ruby dependencies for a specific project. rbenv manages ruby verions. gem install bundler installs Bundler. rbenv rehash is run whenever gems that install new commands are added. bundle install installs all the gem dependencies for a project.

How to start rails server

• The Rails server is hosted on port 3000. This port is forwarded from the virtual machine to the host.
• In order to start the server, cd into the project folder and run rails server -b 0.0.0.0. Then, you should be able to load the root page at localhost:3000.

How to deploy

• Deployment works by pulling down the lastest version of the master from Github and deploying it to the main node. This is done with cap production deploy from within the vm.
• Note: you must have the identity file for the server as well as your Github private key added to your ssh agent in order to connect to the box. The server private key is to access to server obviously, and the Github private key is to pull down the most recent copy of master from Github. To ssh into the vagrant machine, you must run vagrant ssh -- -A. This forwards your ssh agent to the vagrant machine. Capistrano (cap) forwards your ssh-agent to the remote server when it tries to connect, hence why it can access the github repo.
• The server should update automatically. If it crashes, you might have to restart the puma server. (TODO: make this more robust)

Master setup

• Open up ports 22, 80, and 443 for outside access
• Clone project repo to home directory of ubuntu user
• Install ruby, nginx, curl, wget: sudo apt-get ruby nginx curl wget build-essential
• Install chef-solo
• Install librarian-chef and librarian-chef install to download cookbooks
• sudo chef-solo -c solo.rb -j node.json
• Start puma server inside of screen session bundle exec puma -b 'unix:///home/ubuntu/apps/se_filesystem/shared/tmp/sockets/puma.sock' -e production --control 'unix:///home/ubuntu/apps/se_filesystem/shared/tmp/sockets/pumactl.sock' -S /home/ubuntu/apps/se_filesystem/shared/tmp/pids/puma.state (Move this into an init script)

API

• POST /file - Arguments: File_Name, File_Attachement; Uploads File_Attachment to service with name File_Name
• GET /files - Arguments: N/A; Returns list of files present in service
• GET /files/file - Arguments: N/A; Returns metadata for [file]
• DELETE /files/file - Arguments: N/A; Removes [file] from service

Database password needs to be exported to the environment using DATABASE_PASWORD environment variable. Password needs to match password for database role.