Skip to main content
Version: Develop

Creating Backups on S3 Compatible Services

Whether you want to backup Elasticsearch to Amazon AWS S3 or any other S3-compatible service, this section will guide you through the required configuration.

Prerequisites

  • URL to your S3 endpoint (AWS, Minio, Wasabi...)
  • access_key
  • secret_key

Configuring the Backup Service

To configure the S3 endpoint as a backup destination, edit the elasticsearch-backup property in the anynines-backup-manager section of the corresponding a9s Data Service deployment manifest. The configuration should look like this:

...
elasticsearch-backup:
aws_access_key_id: <some_access_key>
aws_secret_access_key: <some_secret_key>
container: <your_container>
endpoint: <some.endpoint.s3.domain>
name:
backup: elasticsearch-backup
restore: elasticsearch-restore
provider: AWS
...

In general, you will only need to replace the above values with the keys and values of your S3 service. However, if you happen to come across some difficulties with the default configuration, take a look at the next section.

virtual hosted_style vs. path_style_access

It might be that your container is only available via the path_style_access (Amazon AWS Documentation). If so, you need to add this line to the elasticsearch-backup: configuration:

...
path_style: true
...

Deprecation Warning

The path_style setting is only viable for Elasticsearch 7.x and upwards. Version 6.x or lower does not support this configuration and thus must use virtual hosted_style paths! Additionally, keep in mind that path_style_access is deprecated (please refer to the Amazon AWS Documentation), so you should generally consider reconfiguring your S3 access if you are still using path_style_access.