Installation/Vagrant puppet capistrano

From diaspora* project wiki
Revision as of 18:31, 31 December 2014 by Joebew42 (talk | contribs) (Deploy a development pod)
Jump to: navigation, search


diaspora* -Replica Is a project that combines Vagrant, Puppet and Capistrano to automate the deploy of a POD. This project is under development, if you have some feedback or do you want to contribute to it, feel free to do it. At this time is necessary to do some test with real deploy on some cloud services (e.g. AWS, Digital Ocean, Rackspace, or other), even on non cloud service.

Some advice before you start to use these tools: You must to have installed on your computer Vagrant 1.6 or above, and Capistrano 3.1 or above (to install Capistrano follow instructions below.)

diaspora* -Replica

The aim of this project is to provide some tools that can help you to deploy a full diaspora* environment through the automation of two tasks:

With these two tasks you can automatically set up different environments, from development to production installation.

Configure FQDNs in your system

Vagrantfile, Puppet and Capistrano are already configured to handle three kind of environment: development, staging and production. If you want to try them you have to update /etc/hosts file, adding to it the three FQDN for the local diaspora* installation.

Put these entries in your /etc/hosts    development.diaspora.local    staging.diaspora.local    production.diaspora.local

Initialize project

git clone
cd diaspora_replica
git submodule update --init

Deploy a development pod

If you are a developer and you want to try diaspora without messing up your computer by installing and configuring extra packages, you can set up a virtual machine that is executed by Vagrant and then automatically configured by Puppet. Now that you have a fully configured virtual machine ready to host a diaspora application, will be very easy to deploy it with Capistrano.

Set up the virtual machine with Vagrant/Puppet

vagrant up development

Wait until the virtual machine is automatically setted up with puppet and is up and running.

Install Capistrano with bundle (if you haven't)

If you have not installed Capistrano on your computer, you can easily run bundle to install it.

cd capistrano/ && bundle

Deploy diaspora* with Capistrano

When the virtual machine is up and running, then you can deploy diaspora* on it using Capistrano

cd capistrano
cap development deploy

When executed the first time, this step can take several minutes (about 20, based on your internet connection), because the diaspora git repository must be cloned. Once capistrano completed the deploy task, you can start diaspora through ``foreman``

cap development foreman:start

Wait until unicorn workers are ready (about 30 seconds) and then your diaspora* installation will be up and running at http://development.diaspora.local

Start, stop and restart

You can use Capistrano tasks to start, stop or restart diaspora*

cap development foreman:start
cap development foreman:stop
cap development foreman:restart

Simulate a production deploy

Simply execute

vagrant up production

and proceed to deploy diaspora* with capistrano:

cd capistrano
cap production deploy deploy:restart

Deploy diaspora* on production environment

If you want to use these tools to deploy a production environment (e.g. stage or production), you have to configure some properties inside Vagrantfile, puppet/manifests/site.pp, capistrano/config/deploy/production.rb and of course, SSL certs and private/public keys for the server.


You have to configure your Vagrantfile based on the virtual machine provider you are going to use (e.g. Amazon AWS, DigitalOcean, and other). Please see the Vagrant Provider Documentation for detailed instructions. If you are not going to use vagrant you can skip this section and apply puppet manually, or configure a puppet master/agent environment. See the Puppet documentation for more informations.


node '' {
  class { 'diaspora':
    hostname            => $fqdn,
    environment         => 'production',
    rvm_version         => '1.26.3',
    ruby_version        => '2.1.5',
    app_directory       => '/home/diaspora',
    user                => 'diaspora',
    group               => 'diaspora',
    db_provider         => 'mysql',
    db_host             => 'localhost',
    db_port             => '3306',
    db_name             => 'diaspora_development',
    db_username         => 'diaspora',
    db_password         => 'diaspora',
    db_root_password    => 'diaspora_root',
    unicorn_worker      => 4,
    sidekiq_concurrency => 5,
    sidekiq_retry       => 10,
    sidekiq_namespace   => 'diaspora',
    foreman_web         => 1,
    foreman_sidekiq     => 1

Of course, you have to change ** with your real Fully Qualified Domain Name, and set up strong password.

Setup the SSL certificate for your server

You have to put the SSL key and certificate in puppet/modules/diaspora/files/certs/. The file names must contain the FQDN followed by .crt and .key. See the examples that already exists.

Setup the public key of the user

Put in puppet/modules/diaspora/files/ the public key of the user that will be granted to execute commands from Capistrano.

Apply Puppet configuration

Now that your Puppet configuration is complete, you have to execute it to your production server. If you use vagrant configured with one of the supported providers it can be done automatically. If you are not able to configure vagrant, you can apply puppet in other ways. But this topic will be not covered here. See the Puppet documentation for this.


Here you have to configure the FQDN, the git repository URL, the name of the branch and the user of the remote server.

Capistrano public key

In order to allow Capistrano to execute commands on the remote server you need to put in capistrano/ssh_keys the private and the public keys of the user. The public key should be the same of puppet/modules/diaspora/files/

Deploy diaspora* with Capistrano

Once you have successfully configured the server, you can deploy and start diaspora*

cd capistrano
cap production deploy deploy:restart

Using PostgreSQL Database

If you want to use PostgreSQL [1] instead of the default MySQL, you can configure it through puppet/manifests/site.pp:

node 'development.diaspora.local' {
  class { 'diaspora':
    hostname         => $fqdn,
    environment      => 'development',
    rvm_version      => '1.25.14',
    ruby_version     => '2.0.0',
    app_directory    => '/home/diaspora',
    user             => 'diaspora',
    group            => 'diaspora',
    db_provider      => 'postgres',
    db_host          => 'localhost',
    db_port          => '5432',
    db_name          => 'diaspora_development',
    db_username      => 'diaspora',
    db_password      => 'diaspora',
    db_root_password => 'diaspora_root'

note the db_provider and db_port parameters.

And you have to uncomment the line:

# set :default_env, { DB: 'postgres' }

That is present in your capistrano/config/deploy/development.rb

[1] Puppet will install PostgreSQL 9.1

PostgreSQL for Staging/Production environments

Because of "--deployment" flag that is set up by default in capistrano bundler, it is necessary to fork diaspora* in a personal git repository and bundle it with PostgreSQL support:

$ DB=postgres bundle

and then add the generated Gemfile.lock under version control. Once you have done that, to enable PostgreSQL you have to uncomment this line:

# set :default_env, { DB: 'postgres' }

in capistrano/config/deploy/production.rb (or capistrano/config/deploy/staging.rb, depends on which environment you are going to deploy.) Of course, you have to specify your git repository, too.

How to set up a Development Environment

You can use these tools to easly set up a fully development environment for diaspora*. The development machine is configured within Vagrantfile with enough RAM (2GB) to run all tests. In this way you can write code using your preferred IDE or editor (vim, emacs, eclipse and so on) directly from your local environment (the host machine), by executing tests within the development virtual machine.

Initialize project

git clone
cd diaspora_replica/
git submodule update --init

Cloning your git repository in src/ directory

Vagrantfile is configured to sync an host directory (src) with a guest directory (diaspora_src). For better I/O performance read the Vagrant Synced Folder Documentation. The first step is to clone your own diaspora* git repository into the local directory src.

cd diaspora_replica/
git clone your_own_diapora_git_repo src

Run development virtual machine

vagrant up development

Enabling vagrant user to use rvm (first time set up)

vagrant ssh development
vagrant@development:~$ sudo usermod -aG rvm vagrant && newgrp rvm
vagrant@development:~$ /bin/bash --login

Prepare the Rails application

vagrant@development:~$ cd diaspora_src

Prepare your configuration files diaspora.yml and database.yml, put it into config/ directory.

Configure Rubies and Gemsets

vagrant@development:~$ rvm use 2.1
vagrant@development:~$ rvm gemset create diaspora_dev
vagrant@development:~$ rvm gemset use diaspora_dev

Install gems and create databases

vagrant@development:~$ bundle
vagrant@development:~$ rake db:create
vagrant@development:~$ rake db:migrate

Run all tests

vagrant@development:~$ rake spec

How to upgrade diaspora*-replica

cd diaspora-replica
git pull --rebase origin master
git submodule update

Which Operating Systems are supported?

  • Ubuntu 14.04LTS Server
  • Ubuntu 12.04LTS Server
  • CentOS 6.4

Choosing the OS from Vagrantfile

By default the Vagrantfile is configured to run an Ubuntu 14.04LTS Server box. If you want to switch to Ubuntu 12.04LTS Server or CentOS 6.4, you have to edit the file.

How to contribute this project

This project is under development. There are a lot of things to do. At the moment the Puppet provides support and, has been tested on Ubuntu 14.04LTS Server, Ubuntu 12.04LTS Server and CentOS 6.4. It could be useful if someone can test it over other version of Ubuntu, or better, can provide support for other distributions (e.g. CentOS). The Database section of the Puppet does not consider parameters like hostname and port at the moment. Furthermore there a lot of variables of diaspora.yml that are not covered (e.g. mail server configuration, unicorn workers, and more).