I've relaunched on Drupal 8.

By Kevin, June 17th, 2019

After many, many years of running on various static site platforms, I've relaunched my blog on Drupal 8.


After returning home from vacation in the Virgin Islands, I noticed my site had an old email address in the metadata of the site. For 5 years I never noticed this. It was really annoying to see and even harder to find where it was coming from. So I decided to make the switch.

For a long time, Jekyll + GitHub Pages worked well enough. But, having to use Vim to edit files, editing by hand, not having the ability to create/work 301 redirects, and general lack of overall control of my site wore me down. In fact, one of the reasons I liked being on GitHub Pages for a static blog was to give me practice learning and putting Vim into practice for text editing - it was a great teaching tool and I still use Vim when ssh'd into servers. But when you want to get down and just write some content, it really gets in the way.

Why Drupal instead of (platform/static)?

Nothing else really suits me. I know the platform, I have total control, I can structure things the way I want and experiment with techniques or features and incorporate them into my projects and vice versa.

How did you migrate?

I employed the use of the Migrate API of Drupal, created 3 migrations (articles, tags, redirects) and pointed all of them at my old websites atom.xml feed. The XML file gave me the entire site, and the XML parser of the Migrate module made the move effortless.

Here is the migration for articles, very simple:

id: article
label: Articles from my Jekyll based blog.
migration_group: content
  plugin: url
  data_fetcher_plugin: http
  data_parser_plugin: simple_xml
  urls: http://kevinquillen.com/atom.xml
    atom: http://www.w3.org/2005/Atom
  item_selector: //atom:entry
      name: id
      label: 'ID'
      selector: atom:id
      name: title
      label: 'Title'
      selector: atom:title
      name: published
      label: 'Published'
      selector: atom:updated
      name: content
      label: 'Content'
      selector: atom:content
      name: tags
      label: 'Tags'
      selector: atom:categories/atom:category
      type: string
  plugin: entity:node
    plugin: default_value
    default_value: article
  title: title
    plugin: format_date
    from_format: 'Y-m-d\TH:i:sP'
    to_format: 'U'
    source: published
    plugin: format_date
    from_format: 'Y-m-d\TH:i:sP'
    to_format: 'U'
    source: published
  'body/value': content
    plugin: default_value
    default_value: full_html
    - plugin: deepen
      source: tags
    - plugin: sub_process
          plugin: migration_lookup
          migration: tags
          source: value

How are you hosting it?

I am using a small Amazon EC2 server and Amazon RDS MySQL database. PHP 7.2 and Memcache. I set the instance up myself. Due to my experiences running all our projects and experiments in Docker containers and general knowledge of Linux, this was not hard to do. I did not go to Acquia, Pantheon or similar managed hosts - the costs are too high for this (~$100-150 a month vs $10), and I wanted total control of my environment.

I am very satisfied with the result.


This was simple. Just like our projects, I set up a private GitLab repo and took advantage of GitLab CI to build the site, run Composer, and deploy that to my Amazon instance automatically. Every merge into the master branch will create a release and deploy it automatically. I love it.

But you're using the Bartik theme!

Yes, I know. Honestly I spent hours looking for themes in various places and nothing suited. I decided to just launch on Bartik - I knew that if I delayed launching when I was so close to the finish line, I would never get it done. Besides, content is king, as long as I can capture the data, I can theme it any way I want when I am ready to do so. I can make those updates over time, and when I know what I want to do, I can create my own theme easily and go from there.