AWS S3 File Adapter for eZ Platform v3

This bundle provides AWS S3 file adapter, and the key features are:

• Easy to install and configure.
• Each environment uses its own storage within the same AWS S3 Bucket. And the bundle provides the script to sync storage between different environments.
• Direct AWS S3 links are used for generated image aliases.
• It provides an additional script to generate specific image aliases for specific content items. Which might be used as a performance optimization in some cases.
• Binary files are uploaded with private ACL. So it is impossible to access them bypassing Ibexa DXP security rules.

Installation

1. Using AWS S3 create a new public bucket.

2. Using AWS IAM create a new "Programmatic access" user with read/write access to the previously created S3 bucket.

3. Set AWS credentials as environment variables. Run the following command, if the project runs on Ibexa Cloud or Platform.sh:

   platform variable:create --level=project --name=AWS_ACCESS_KEY_ID \
       --value=<YOUR_KEY_ID> --json=false --sensitive=false --prefix=env \
       --visible-build=false --visible-runtime=true --enabled=true --inheritable=true
   platform variable:create --level=project --name=AWS_SECRET_ACCESS_KEY \
       --value=<YOUR_ACCESS_KEY> --json=false --sensitive=true --prefix=env \
       --visible-build=false --visible-runtime=true --enabled=true --inheritable=true

4. Add variables to .env file:

   ...
   ###> contextualcode/ezplatform-aws-s3-adapter ###
   AWS_FILE_ADAPTER_ACCESS_KEY_ID=${AWS_ACCESS_KEY_ID}
   AWS_FILE_ADAPTER_SECRET_ACCESS_KEY=${AWS_SECRET_ACCESS_KEY}
   AWS_FILE_ADAPTER_REGION=us-east-1
   AWS_FILE_ADAPTER_BUCKET=<S3_BUCKET_NAME>
   AWS_FILE_ADAPTER_PREFIX=${PLATFORM_BRANCH}
   ###< contextualcode/ezplatform-aws-s3-adapter ###

5. Using composer require contextualcode/ezplatform-aws-s3-adapter:

   composer require contextualcode/ezplatform-aws-s3-adapter

6. Update config/packages/oneup_flysystem.yaml:

   oneup_flysystem:
       adapters:
           aws_s3_adapter:
               awss3v3:
                   client: 'amazon.client.aws_s3_adapter'
                   bucket: '%env(AWS_FILE_ADAPTER_BUCKET)%'
                   prefix: '%env(AWS_FILE_ADAPTER_PREFIX)%'
   
   ez_io:
       binarydata_handlers:
           aws_s3:
               flysystem:
                   adapter: aws_s3_adapter
       metadata_handlers:
           dfs:
               legacy_dfs_cluster:
                   connection: doctrine.dbal.default_connection
   
   ezplatform:
       system:
           default:
               io:
                   binarydata_handler: aws_s3
                   metadata_handler: dfs

7. Migrate the existing files if you are enabing AWS S3 File Adapter on the existing project:

   php -d memory_limit=-1 bin/console ezplatform:io:migrate-files --from=default,default --to=dfs,aws_s3 --no-interaction

8. Deploy the changes:

   git add composer.json composer.lock symfony.lock .env config/bundles.php config/packages/oneup_flysystem.yaml
   git commit -m "Installed AWS S3 file adapter"
   git push

Syncing storage between different environments

This bundle provides a script to sync storage between different environments. This script depends on AWS CLI. So please make sure it is installed and set up. Usage:

bin/sync-environment-storage <AWS_S3_BUCKET_NAME> <DESTINATION_ENVIRONMENT> <SOURCE_ENVIRONMENT>

So if you want to sync the storage from the master environment into the staging environment for the project which uses my-project-storage AWS S3 bucket you should run:

bin/sync-environment-storage my-project-storage staging master

Generating image aliases

In addition, this bundle provides bin/regenerate-image-aliases script. Which allows regenerating image aliases for specific content. Usage:

bin/regenerate-image-aliases <CLEAR_BEFORE> <LOCATION_IDS> <CONTENT_TYPES> <ALIASES>

To regenerate "medium" and "large" aliases created in the last 24 hours for all folders and articles located under root location (location id: 1), please run:

bin/regenerate-image-aliases '24 hours ago' '1' 'folder,article' 'medium,large'

Depending on the project needs, a similar script might be added to a cron job. So when the page will be rendered its image aliases will be already generated, and no CPU resources are used for image aliases generation.