Configuring CloudMirror replication

The StorageGRID Webscale CloudMirror replication service enables a tenant to automatically replicate objects to an external S3 bucket. You can enable replication using the Tenant Manager.

Before you begin

About this task

CloudMirror replication copies objects from a source bucket to a destination bucket that is specified in an endpoint. To enable CloudMirror replication for a bucket, you must create and apply valid bucket replication configuration XML. The replication configuration XML must use the URN of an S3 bucket endpoint for each destination.

For general information on bucket replication and how to configure it, see the Amazon documentation on Cross-Region replication. For information on how StorageGRID Webscale implements the S3 bucket replication configuration API, see the instructions for implementing S3 client applications.

If you enable CloudMirror replication on a bucket that contains objects, new objects added to the bucket are replicated, but the existing objects in the bucket are not. You must update existing objects to trigger replication.

If you specify a storage class in the replication configuration XML, StorageGRID Webscale uses that class when performing operations against the destination S3 endpoint. The destination endpoint must also support the specified storage class. Be sure to follow any recommendations provided by the destination system vendor.

Steps

  1. Enable replication for your source bucket:
    1. Use a text editor to create the replication configuration XML required to enable replication, as specified in the S3 replication API.
      When configuring the XML:
      • Use the URN of an S3 bucket endpoint as the destination.
      • Optionally add the <StorageClass> element, and specify one of the following:
        • STANDARD: The default storage class. If you do not specify a storage class when you upload an object, the STANDARD storage class is used.
        • STANDARD_IA: (Standard - infrequent access.) Use this storage class for data that is accessed less frequently, but that still requires rapid access when needed.
        • REDUCED_REDUNDANCY: Use this storage class for noncritical, reproducible data that can be stored with less redundancy than the STANDARD storage class.
      Example
      <ReplicationConfiguration>
          <Role></Role>
          <Rule>
              <Status>Enabled</Status>
              <Prefix>2017</Prefix>
              <Destination>
                 <Bucket>urn:sgws:s3:::2017-records</Bucket>
                 <StorageClass>STANDARD</StorageClass>
               </Destination>
           </Rule>
      </ReplicationConfiguration>
    2. In the Tenant Manager go to S3 > Buckets.
    3. Select the source bucket, then click Configure Replication.
    4. Paste the replication configuration into the text box, and click Save.

      screenshot of replication configuration dialog box
      Note: Platform services must be enabled for each tenant account by a StorageGRID Webscale administrator using the Grid Manager or Management API. Contact your grid administrator if an error occurs when you save the configuration XML.
  2. Verify that replication is configured correctly:
    1. Add an object to the source bucket that meets the requirements for replication as specified in the replication configuration.
      In the example shown earlier, objects that match the prefix "2017" are replicated.
    2. Confirm that the object has been replicated to the destination bucket.
      For small objects, replication happens quickly.
    You have configured your source bucket for StorageGRID Webscale bucket replication.