Here we describe how to install OpenDSA-LTI on a single Ubuntu Server 14.04.3 LTS 64-bit from scratch.

Hardware requirements

The following server requirements will be fine for supporting hundreds of users.

• Ubuntu Server 14.04.3 LTS
• Minimum 8GB of memory
• At least two 2.00GHz CPU
• Minimum 250GB of free disk

Installation Instructions

Install OpenDSA-DevStack on your local machine

• OpenDSA-LTI is using a remote server automation and deployment tool called Capistrano. Each time you want to deploy new changes to the OpenDSA-LTI production server you have to initiate the deployment command from within the development environment OpenDSA-DevStack.

• Once you have OpenDSA-DevStack up and running open a new terminal and do the following to generate a pair of authentication keys. Note: Do not enter a passphrase.

$ cd OpenDSA-DevStack
$ vagrant ssh
$ ssh-keygen -t rsa

• You will see something similar to the following

Generating public/private rsa key pair.
Enter file in which to save the key (/home/vagrant/.ssh/id_rsa):
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in /home/vagrant/.ssh/id_rsa.
Your public key has been saved in /home/vagrant/.ssh/id_rsa.pub.
The key fingerprint is:
00:54:14:9d:c2:3d:d8:5a:d0:12:ae:0c:d6:09:e6:88 vagrant@vagrant-ubuntu-trusty-64
The key's randomart image is:
+--[ RSA 2048]----+
|  o.o+*X .       |
|.+ o += x        |
|E + o o= .       |
| . o ...         |
|    o   S        |
|                 |
|                 |
|                 |
|                 |
+-----------------+

Creating deploy user on your production server

• The first thing we will do on our new server is to create the user account we'll be using to run OpenDSA-LTI and work from there. Open a new terminal, ssh to your production server, and do the following

$ sudo adduser deploy
$ sudo adduser deploy sudo
$ su deploy
$ cd
$ mkdir .ssh
$ cd .ssh
$ touch authorized_keys

• Before we move forward is that we're going to setup SSH to authenticate via keys instead of having to use a password to login. It's more secure and will save you time in the long run. Switch back to OpenDSA-DevStack terminal to append the new public key to deploy@<prod_server>:.ssh/authorized_keys and enter deploy user password one last time:

$ cd
$ cat .ssh/id_rsa.pub | ssh deploy@<prod_server> 'cat >> .ssh/authorized_keys'

• For now you can ssh to you production server from within OpenDSA-DevStack without password

$ ssh deploy@<prod_server>

• For the next steps, make sure you are logged in as the deploy user on the production server!

Installing Ruby

• The first step is to install some dependencies for Ruby.

$ sudo apt-get update
$ sudo apt-get install -y git-core curl zlib1g-dev build-essential libssl-dev libreadline-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt1-dev libcurl4-openssl-dev python-software-properties libffi-dev dkms libxslt-dev libpq-dev python-dev python-pip python-feedvalidator python-sphinx libmariadbclient-dev libevent-dev libsqlite3-dev

• Next we're going to be installing Ruby using rbenv.

$ cd
$ git clone https://github.com/rbenv/rbenv.git ~/.rbenv
$ echo 'export PATH="$HOME/.rbenv/bin:$PATH"' >> ~/.bashrc
$ echo 'eval "$(rbenv init -)"' >> ~/.bashrc
$ exec $SHELL

$ git clone https://github.com/rbenv/ruby-build.git ~/.rbenv/plugins/ruby-build
$ echo 'export PATH="$HOME/.rbenv/plugins/ruby-build/bin:$PATH"' >> ~/.bashrc
exec $SHELL

$ rbenv install 2.3.1
$ rbenv global 2.3.1
$ ruby -v

• The last step is to install Bundler

$ gem install bundler

• Run rbenv rehash after installing bundler.

Installing Nginx

• Phusion is the company that develops Passenger and they recently put out an official Ubuntu package that ships with Nginx and Passenger pre-installed. We'll be using that to setup our production server because it's very easy to setup.

$ sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys 561F9B9CAC40B2F7
$ sudo apt-get install -y apt-transport-https ca-certificates

• Add Passenger APT repository

$ sudo sh -c 'echo deb https://oss-binaries.phusionpassenger.com/apt/passenger trusty main > /etc/apt/sources.list.d/passenger.list'
$ sudo apt-get update

• Install Passenger & Nginx

$ sudo apt-get install -y nginx-extras passenger

• So now we have Nginx and passenger installed. We can manage the Nginx webserver by using the service command:

$ sudo service nginx start

• Open up the server's IP address in your browser to make sure that nginx is up and running. The service command also provides some other methods such as restart and stop that allow you to easily restart and stop your webserver.
• Next, we need to update the Nginx configuration file and update a couple of things. You'll want to open up /etc/nginx/nginx.conf in your favorite editor or simply use nano:

$ sudo nano /etc/nginx/nginx.conf

• First, change user from www-data to deploy

user `deploy`;
worker_processes auto;
pid /run/nginx.pid;

events {
        worker_connections 768;
        # multi_accept on;
}
...

• Second, point Passenger to the version of Ruby that we're using. Find the following lines in the configuration file

##
# Phusion Passenger
##
# Uncomment it if you installed ruby-passenger or ruby-passenger-enterprise
##

• Then put the following two lines right after them

passenger_root /usr/lib/ruby/vendor_ruby/phusion_passenger/locations.ini;
passenger_ruby /home/deploy/.rbenv/shims/ruby;

• The passenger_ruby is the important line here. Once you've changed passenger_ruby to use the right version Ruby, you can restart Nginx with the new Passenger configuration.

$ sudo service nginx start

• Now that we've restarted Nginx, the OpenDSA-LTI will be served up using the deploy user. We can make sure that our web server will restart automatically when the server is rebooted by typing:

$ sudo update-rc.d nginx defaults

• This should already be enabled by default, so you may see a message like this:

$ System start/stop links for /etc/init.d/nginx already exist.

• This just means that it was already configured correctly and that no action was necessary. Either way, your Nginx service is now configured to start up at boot time.

Setting Up MySQL server

• You can install MySQL server and client from the packages in the Ubuntu repository. As part of the installation process, you'll set the password for the root user. This information will go into your OpenDSA-LTI database.yml file in the future.

$ sudo apt-get install -y mysql-server mysql-client libmysqlclient-dev

• Installing the libmysqlclient-dev gives you the necessary files to compile the mysql2 gem which is what Rails will use to connect to MySQL when you setup OpenDSA-LTI application.

• Now we will create a new database and user opendsa for OpenDSA-LTI application. First login to mysql

$ mysql -uroot -p

• Then create opendsa database

CREATE DATABASE opendsa DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;

GRANT ALL PRIVILEGES ON opendsa.* to 'opendsa'@'localhost'  IDENTIFIED BY '<db_password>';
FLUSH PRIVILEGES;
exit

Install Node.js and bower

• Node.js is required by Rails assets pipeline.

$ sudo apt-get install -y nodejs
$ sudo ln -s /usr/bin/nodejs /usr/sbin/node
$ sudo npm install -g jshint
$ sudo npm install -g csslint
$ sudo npm install -g bower

Clone OpenDSA repository in your production server

• OpenDSA contains all the book contents that will be served by OpenDSA-LTI Rails application. You only need to clone OpenDSA under deploy home directory, Then all the linking between OpenDSA and OpenDSA-LTI will happen automatically through the automated deployment tasks.

$ cd
$ git clone https://github.com/OpenDSA/OpenDSA.git
$ cd OpenDSA
$ make pull

• For the next steps, Switch back to OpenDSA-DevStack terminal

Deploy OpenDSA-LTI

• You need to make some changes to OpenDSA-LTI repository related to your specific production server. To do that you need to fork OpenDSA-LTI to your GitHub account and then add your own repository as a remote to OpenDSA-LTI in OpenDSA-DevStack. This way you can make your own changes to OpenDSA-LTI and keep up to date with the latest changes done in the originial repository.
• In your OpenDSA-DevStack terminal, add your forked repository, replace your_username with you github account

$ cd /vagrant/OpenDSA-LTI
$ git remote add forked https://github.com/your_username/OpenDSA-LTI.git

• First, in /vagrant/OpenDSA-LTI/config/deploy.rb file, change repo_url to match you forked repository url. Again replace your_username with you github account

# config valid only for Capistrano 3.1
lock '3.2.1'

set :application, 'OpenDSA-LTI'
# set :repo_url, 'git://github.com/OpenDSA/OpenDSA-LTI.git'
set :repo_url, 'git://github.com/your_username/OpenDSA-LTI.git'


# Default branch is :master
# ask :branch, proc { `git rev-parse --abbrev-ref HEAD`.chomp }.call
...

• Second, you need to open up /vagrant/OpenDSA-LTI/config/deploy/production.rb file to set the server IP address of your production server. Replace 128.173.236.80 with your server IP address.

set :stage, :production

# Simple Role Syntax
# ==================
# Supports bulk-adding hosts to roles, the primary server in each group
# is considered to be the first unless any hosts have the primary
# property set.  Don't declare `role :all`, it's a meta role.

role :app, %w{[email protected]}
role :web, %w{[email protected]}
role :db,  %w{[email protected]}


# Extended Server Syntax
# ======================
# This can be used to drop a more detailed server definition into the
# server list. The second argument is a, or duck-types, Hash and is
# used to set extended properties on the server.
server '128.173.236.80', user: 'deploy', roles: %w{web app db}, my_property: :my_value
...

• Finally, if you have staging server you need to open up /vagrant/OpenDSA-LTI/config/deploy/staging.rb file to set the server IP address of you staging server. Replace 128.173.236.221 with your staging server IP address.

• Commit your changes and push it to forked remote

$ cd /vagrant/OpenDSA-LTI
$ git add .
$ git commit -m "Add production server details"
$ git push forked master

• Switch to OpenDSA-DevStack terminal
• Deploy OpenDSA-LTI for the first time (this step will fail!). But it will create OpenDSA-LTI folder structure in the production server.

$ cd /vagrant/OpenDSA-LTI
$ bundle exec cap production deploy

• First-time deployment failed because the shared files database.yaml and secrets.yaml weren't created on the production server yet.
• Switch back to the production server terminal and do the following

$ cd /home/deploy/OpenDSA-LTI/shared/config
$ touch database.yml
$ touch secrets.yml

• You need to put database credentials in database.yml and generate a new secret for production and save it in secrets.yml.
• First update databse.yml

$ cd /home/deploy/OpenDSA-LTI/shared/config
$ sudo nano database.yml

• Copy the following lines and replace db_password with your password

production:
   adapter: mysql2
   database: opendsa
   username: opendsa
   password: db_password
   host: localhost
   strict: false

staging:
   adapter: mysql2
   database: opendsa
   username: opendsa
   password: db_password
   host: localhost
   strict: false

• Second update secrets.yml. To generate a new secret go to OpenDSA-DevStack terminal

$ cd /vagrant/OpenDSA-LTI
$ rake secret

• A new secret will be generated, copy that string and past it in secrets.yml file on the production server. Open secrets.yml file

$ cd /home/deploy/OpenDSA-LTI/shared/config
$ sudo nano secrets.yml

• Copy the following lines and replace secret_string with your new secret

production:
  secret_key_base: secret_string

staging:
  secret_key_base: secret_string

• Now production server is ready for deployment, switch back to OpenDSA-DevStack terminal and execute the following

$ cd /vagrant/OpenDSA-LTI
$ bundle exec cap production deploy

• If you have configured a staging server you can deploy your changes to the staging server the same way as the production. Switch to OpenDSA-DevStack terminal and execute the following

$ cd /vagrant/OpenDSA-LTI
$ bundle exec cap staging deploy

Final Steps

• Switch to the production server terminal
• Adding The Nginx Host. In order to get Nginx to respond with the Rails application, we need to modify it's sites-enabled. Open up /etc/nginx/sites-enabled/default in your text editor and we will replace the file's contents with the below configuration. Replace prod_server_name with your domain name.

server {
        listen 80 default_server;
        listen [::]:80 default_server ipv6only=on;

        listen 443 ssl;

        server_name prod_server_name;
        ssl_certificate /etc/nginx/ssl/nginx.crt;
        ssl_certificate_key /etc/nginx/ssl/nginx.key;

        passenger_enabled on;
        rails_env    production;
        root         /home/deploy/OpenDSA-LTI/current/public;

        # redirect server error pages to the static page /50x.html
        error_page   500 502 503 504  /50x.html;
        location = /50x.html {
            root   html;
        }
}

• Replace /etc/nginx/ssl/nginx.crt and /etc/nginx/ssl/nginx.key with your valid certificate and key.

• Restart Nginx web server

$ sudo service nginx restart

• Go to https://prod_server_name you should see OpenDSA Rails application landing page. Congratulations!!!

Production deployment workflow

• Follow the instructions on the OpenDSA-DevStack page to perform a deployment on the production server.

Export anonymized OpenDSA-LTI data

• OpenDSA stores massive interactions data and exercises attempts. If you want to use this data for learning analytics research tasks, you may need to export the data from the database and anonymize it first.

Export and anonymize

• Export databse schema from https://opendsa-server.cs.vt.edu server using MySQL workbench data export tool
• Provision the OpenDSA-DevStack VM
• vagrant ssh into the VM
• cd /vagrant/OpenDSA-LTI
• bundle exec rake db:drop
• bundle exec rake db:create
• Connect to your local OpenDSA-DevStack using MySQL workbench and import the data using data import tool
• On OpenDSA-DevStack database do the follwoing:
• Relax the following constraints
  • alter table users drop index email;
  • alter table users drop index slug;
• Change the following configuration on workbench
  • Edit > Preferences > Sql Editor > uncheck the "Safe Updates"
• Execute the following update statements to anonymize the exported data
  • UPDATE opendsa.users SET email = "[email protected]";
  • UPDATE opendsa.users SET slug = "[email protected]";
  • UPDATE opendsa.users SET first_name = "first_name";
  • UPDATE opendsa.users SET last_name = "last_name";
  • UPDATE opendsa.users SET encrypted_password = "encrypted_password";
• Re-export the anynomized schema from OpenDSA-DevStack.

Import anonymized data

• Provision the OpenDSA-DevStack VM
• vagrant ssh into the VM
• cd /vagrant/OpenDSA-LTI
• bundle exec rake db:drop
• bundle exec rake db:create
• Connect to your local OpenDSA-DevStack using Mysql workbench and import the anonymized data using data import tool