Jekyll S3 Deploy Command

Publish your Jekyll site to Amazon S3. This is a pure-ruby solution that uses on aws-sdk, as opposed to s3_website which relies on a Java JAR.

Follow these instructions to setup your Amazon AWS bucket as appropriate. You also probably want to setup your own domain although this process has no bearing on the actual upload/ serving files from S3.

Installation and Usage

Add this line to your Jekyll project's Gemfile:

group :jekyll_plugins do
  gem 'jekyll-s3_deploy', github: 'thom-nic/jekyll-s3_deploy'
end

Then run bundle install. Add a _config_s3.yml file with the following:

# S3 access credentials
# Add this file to your .gitignore!!!
access_key: your access key
secret_key: your secret key

And add the following to _config.yml

s3:
  bucket: mysite.com
  region: us-east-1   # optional, this is the default value
  cache_control:      # Cache-Control headers (optional)
    'assets/*' : max-age=3600
    '*.html' : max-age=300
  deflate:            # what assets should be Content-Encoded (optional)
    - '*.html'
    - '*.css'
    - '*.svg'

And then execute:

jekyll deploy_s3

If you use the Jekyll sitemap plugin, this project also contains a ping command to notify Google and Bing web crawlers that your site has been updated. Just run jekyll ping

Options

-f - Force upload everything under _site

More details on setting up your AWS credentials...

This is the one bit that's difficult to remember/ not explained well in the AWS docs. When you create your IAM user, this is the policy you need:

{
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "s3:ListAllMyBuckets",
      "Resource": "arn:aws:s3:::*"
    },
    {
      "Effect": "Allow",
      "Action": "s3:*",
      "Resource": [
        "arn:aws:s3:::mysite.com",
        "arn:aws:s3:::mysite.com/*"
      ]
    }
  ]
}

(of course, replace mysite.com with your domain)

To Do

Document command options
Add CloudFront support
Add gzip content-encoding support