In this post I will show you how to build the site every time you push something to the repository and push it automatically to S3. I use Travis Ci to do the automted build and upload to S3.
To use Travis CI you have to add a .travis.yaml in the root of your repository. The file is the configuration for our build. Here my configuration to build, test and deploy my websites.
# Auto deploy repo from Github to Amazon S3 bucket via Travis CI # * Set env vars for ACCESS_KEY_ID, BUCKET_NAME and SECRET_ACCESS_KEY on Travis # * Update `bucket.name` in `sync` command # * Assumes your `publishDir` is the default (`public`) - if not update `sync` command language: go before_install: - export NOKOGIRI_USE_SYSTEM_LIBRARIES=true addons: apt: packages: - libcurl4-openssl-dev # required to avoid SSL errors install: go get -v github.com/spf13/hugo before_script: - ruby --version - gem install html-proofer - python --version - sudo pip install s3cmd script: - hugo - htmlproofer public/ after_success: - s3cmd sync --delete-removed --no-preserve --acl-private --access_key=$ACCESS_KEY_ID --secret_key=$SECRET_ACCESS_KEY -r public/ s3://$BUCKET_NAME notifications: email: true
Up to the script block travis is installing some dependencies.
In the script block do the real work by generate the page and test the page with the html-proffer tool.
If the script block run without problem it will execute the after_script block. In the blog travis deploy the generated website to S3.
Setup your repository for travis
To enable your website repo you can just enable it in the webinterface from Travis CI.
My script also need to set up three following environment variables: * ACCESS_KEY_ID (aws access key) * SECRET_ACCESS_KEY (aws secrect access key) * BUCKET_NAME (bucket name to deploy to)
Just setup the variables in the build settings webinterface.
I deploy all files with the option –acl-private this set the ACL the only I can access the files. How to allow the access over cloudflare I write here.