jekyll-hook

Note: This tool is no longer maintained. There are a number of newer and better approaches to update jekyll-powered websites. We use Continious Integration (e.g. Travis) as a replacement for jekyll-hook at Development Seed. Read more here: https://jekyllrb.com/docs/deployment-methods/

A server that listens for webhook posts from GitHub, generates a website with Jekyll, and moves it somewhere to be published. Use this to run your own GitHub Pages-style web server. Great for when you need to serve your websites behind a firewall, need extra server-level features like HTTP basic auth (see below for an NGINX config with basic auth), or want to host your site directly on a CDN or file host like S3. It's cutomizable with two user-configurable shell scripts and a config file.

This guide is tested on Ubuntu 14.0

Dependencies Installation

First install main dependencies

$: sudo apt-get update
$: sudo apt-get install git nodejs ruby ruby1.9.1-dev npm

Symlink nodejs to node

$: sudo ln -s /usr/bin/nodejs /usr/bin/node

To keep server running we use Forever:

$: sudo npm install -g forever

We also need Jekyll and Nginx

$: sudo gem install jekyll rdiscount json
$: sudo apt-get install nginx

Installation

Clone the repo

$: git clone https://github.com/developmentseed/jekyll-hook.git

Install dependencies:

$: cd jekyll-hook
jekyll-hook $: npm install

If you receive an error similar to this npm ERR! Error: EACCES, mkdir '/home/ubuntu/tmp/npm-2223-4myn3niN' run:

$: sudo chown -R ubuntu:ubuntu /home/ubuntu/tmp
$: npm install

You should replace ubuntu with your username

Configuration

Copy config.sample.json to config.json in the root directory and customize:

$: cp config.sample.json config.json
$: vim config.json

Configuration attributes:

• gh_server The GitHub server from which to pull code, e.g. github.com
• temp A directory to store code and site files
• public-repo Whether the repo is public or private (default is public)
• scripts
  • [branch-name] (optional)
    • build The build script to run for a specific branch.
    • publish The publish script to run for a specific branch.
  • #default
    • build The build script to run if no match was found for the branch specified in the webhook.
    • publish The publish script to run if match was found for the branch specified in the webhook.
• secret Optional. GitHub webhook secret.
• email Optional. Settings for sending email alerts
  • isActivated If set to true email will be sent after each trigger
  • user Sending email account's user name (e.g. example@gmail.com)
  • password Sending email account's password
  • host SMTP host for sending email account (e.g. smtp.gmail.com)
  • ssl true or false for SSL
• accounts An array of accounts or organizations whose repositories can be used with this server

You can also adjust build.sh and publish.sh to suit your workflow. By default, they generate a site with Jekyll and publish it to an NGINX web directory.

Webhook Setup on Github

Set a Web hook on your GitHub repository that points to your jekyll-hook server http://example.com:8080/hooks/jekyll/:branch, where :branch is the branch you want to publish. Usually this is gh-pages or master for *.github.com / *.github.io repositories.

For every branch you want to build/publish (as defined in the scripts) you need to set up a different webhook.

Configure a webserver (nginx)

The default publish.sh is setup for nginx and copies _site folder to /usr/share/nginx/html/rep_name.

If you would like to copy the website to another location, make sure to update nginx virtual hosts which is located at /etc/nginx/nginx/site-available on Ubuntu 14.

You also need to update publish.sh

For more information Google or read this:

Launch

$: ./jekyll-hook.js

To launch in background run:

$: forever start jekyll-hook.js

To kill or restart the background job:

    $: forever list
    info:    Forever processes running
    data:        uid  command         script         forever pid  logfile                        uptime
    data:    [0] ZQMF /usr/bin/nodejs jekyll-hook.js 4166    4168 /home/ubuntu/.forever/ZQMF.log 0:0:1:22.176
    $: forever stop 0

Publishing content

S3

To publish the site on Amazon S3, you need to install S3cmd. On Ubuntu run:

$: sudo apt-get install s3cmd
$: s3cmd --configure

For more information read this.

scripts/publish-s3.sh does the rest of the job for you. Just make sure to add your bucket name there.

More details on build.sh

The stock build.sh copies rendered site files to subdirectories under a web server's www root directory. For instance, use this script and NGINX with the following configuration file to serve static content behind HTTP basic authentication:

server {
    root /usr/share/nginx/www;
    index index.html index.htm;

    # Make site accessible from http://localhost/
    server_name localhost;

    location / {
        # First attempt to serve request as file, then
        # as directory, then fall back to index.html
        try_files $uri $uri/ /index.html;

        # Optional basic auth restriction
        # auth_basic "Restricted";
        # auth_basic_user_file /etc/nginx/.htpasswd;
    }
}

Replace this script with whatever you need for your particular hosting environment.

You probably want to configure your server to only respond POST requests from GitHub's public IP addresses, found on the webhooks settings page.