Deploying Jekyll on a Linode VM via GitHub Actions

In my GitHub Actions, “Getting your feet wet”, guide series I went over how to build a docker container, automatically updating it using a app called Diun, deploying the container update to a VM in Linode, and creating a releasable binary. All based on the web server I use for my homelab, Caddy. This post was not originally planned to be apart of the series but I am including it as an example of a more, “advanced”, actions deployment. By default Jekyll deploys to GitHub Pages which is really nice if you do not want to spin up your own web server.

For me though I wanted to host it myself. The process for this was mostly straight forward as Jekyll already includes the ability to self host. The diagram below demonstrates the GitHub Actions flow for both the default deployment to pages and my deployment to a Linode VM.

Workflows for deploying Jekyll via GitHub Actions

If you want to use the actions file below you will need the following items completed beforehand.

Cloned / Setup a Jekyll blog in a repository on GitHub.
Have a VM running that is accessible by SSH from the internet (WAN).
- I recommend Linode if you are just starting out.
Have a user account created on the VM for actions to log into via SSH.
- You can check out my post, here, which details this process.
Have a domain registered and pointing to the public (WAN) IP of the VM.
Have a web server installed and setup to serve from the directory the blog will be in.
- I am using Caddy to serve the blog directory.
Repository secrets setup:
- LINODEFINGERPRINT = ssh host fingerprint
- LINODEHOST = hostname/wan ip of the VM
- LINODEUSER = username actions needs for login
- LINODESSHKEY = ssh private key for above user
- DISCORD_WEBHOOK = text channel webhook url. Only needed if you use the notify job.

GitHub Actions workflow for deploying the Jekyll blog:

1
name: 'Deploy Blog'
2
on:
3
  push:
4
    branches:
5
      - main
6
    paths-ignore:
7
      - .gitignore
8
      - README.md
9
      - LICENSE
10
  workflow_dispatch:
11

12
jobs:
13
  # build and create the jekyll blog artifact
14
  # include 'skip ci' in commit message to not run this workflow
15
  build:
16
    if: "github.ref == 'refs/heads/main' && !contains(github.event.head_commit.message, 'skip ci')"
17
    runs-on: ubuntu-latest
18
    steps:
19
      - name: checkout repo
20
        uses: actions/checkout@v2
21
        with:
22
          fetch-depth: 0  # for posts's lastmod
23
      - name: setup ruby
24
        uses: ruby/setup-ruby@v1
25
        with:
26
          ruby-version: 2.7
27
          bundler-cache: true
28
      - name: build production
29
        env:
30
          _config: _config.yml
31
          _baseurl:
32
        run: JEKYLL_ENV=production bundle exec jekyll b --incremental
33
      - name: copy required files
34
        run: |
35
          cp -r _data _site/data
36
          cp -r _plugins _site/plugins
37
          cp -r _tabs _site/tabs
38
          cp _config.yml _site/config.yml
39
      - name: upload blog
40
        uses: actions/upload-artifact@v3
41
        with:
42
          path: _site/*
43
          retention-days: 1
44

45
  # grab the build artifact and test it using htmlproofer
46
  test:
47
    needs: build
48
    runs-on: ubuntu-latest
49
    steps:
50
      - name: download blog
51
        uses: actions/download-artifact@v3
52
        with:
53
          name: artifact
54
          path: _site/
55
      - name: test production site
56
        uses: chabad360/htmlproofer@master
57
        with:
58
          directory: "_site/"
59
          arguments: --disable-external --check-html --allow_hash_href=true
60

61
  # grab artifact if test job is successful and deploy it using rsync
62
  deploy:
63
    needs: test
64
    runs-on: ubuntu-latest
65
    steps:
66
      - name: download blog artifact
67
        uses: actions/download-artifact@v3
68
        with:
69
          name: artifact
70
          path: _site/
71
      - name:
72
        run: |
73
          mkdir ~/.ssh
74
          touch ~/.ssh/known_hosts
75
          echo ${{ secrets.LINODEFINGERPRINT }} >> ~/.ssh/known_hosts
76
      - name: deploy blog
77
        uses: burnett01/rsync-deployments@5.2.1
78
        with:
79
          switches: -avr
80
          path: _site/
81
          remote_path: /opt/sites/blog/
82
          remote_host: ${{ secrets.LINODEHOST }}
83
          remote_port: 22
84
          remote_user: ${{ secrets.LINODEUSER }}
85
          remote_key: ${{ secrets.LINODESSHKEY }}
86

87
  # use nobrayner/discord-webhook action to notify status.
88
  # delete this job if you do not want to use it.
89
  notify:
90
    name: notify
91
    runs-on: ubuntu-latest
92
    needs:
93
      - build
94
      - test
95
      - deploy
96
    if: ${{ always() }}
97
    steps:
98
      - name: send notif discord
99
        uses: nobrayner/discord-webhook@v1
100
        with:
101
          github-token: ${{ secrets.github_token }}
102
          discord-webhook: ${{ secrets.DISCORD_WEBHOOK }}
103
          include-details: 'true'
104
          color-success: '#7ED321'
105
          color-failure: '#D0021B'
106
          color-cancelled: '#9013FE'
107
          avatar-url: 'https://octodex.github.com/images/Terracottocat_Single.png'
108
          title: '${{ github.workflow }}'
109
          description: '${{ github.workflow }} was triggered. It returned status: {{STATUS}}'

Workflow file for deleting artifacts to save GitHub storage space:

1
# trigger workflow to run after listed workflows are run
2
name: 'Delete Old Artifacts'
3
on:
4
  workflow_run:
5
    workflows: [Deploy Blog, Deploy Blog (DEV)]
6
    types:
7
      - completed
8

9
# delete all leftover artifacts since GitHub's lowest retention time period is 24 hours
10
jobs:
11
  delete-artifacts:
12
    runs-on: ubuntu-latest
13
    steps:
14
      - uses: kolpav/purge-artifacts-action@v1
15
        with:
16
          token: ${{ secrets.GITHUB_TOKEN }}
17
          expire-in: 0 # Setting this to 0 will delete all artifacts

Caddyfile example:

1
{
2
        email email@email.com
3
#       acme_ca https://acme-staging-v02.api.letsencrypt.org/directory
4
        debug
5
        log {
6
                output file /var/log/caddy/caddy.log {
7
                        roll_size 12mb
8
                        roll_keep 10
9
                        roll_keep_for 24h
10
                }
11
                format json {
12
                        message_key msg
13
                        level_key debug
14
                        time_key ts
15
                        name_key name
16
                        time_format wall
17
                }
18
        }
19
}
20

21
(tls) {
22
        tls {
23
                dns cloudflare {env.CF_API_KEY}
24
        }
25
}
26

27
blog.alexsguardian.net {
28
        import tls
29
        root * /opt/sites/blog/
30
        file_server
31
        handle_errors {
32
                rewrite * /404.html
33
                file_server
34
        }
35
}

This post is merely an example for the GitHub Actions guide series. I actually use the above workflow to deploy this blog from GitHub, so it does work! If you have any questions, or a suggestion, feel free to reach out via a comment or email!

Affiliate Links

Linode💜

_{FTC: Some links in this post are income generating.}

Deploying Jekyll on a Linode VM via GitHub Actions

Affiliate Links

Comments

Related Posts

GitHub Actions - Automatically Deploy Docker Updates to Cloud Hosts via SSH

GitHub Actions - Automatically Build and Release Binaries

GitHub Actions - Automating your Docker Builds with Diun