AWS S3 Replication and Event Notifications

S3 Replication

We can configure S3 buckets to replicate their data to other buckets in the same region (same-region replication, or SRR) or across regions (cross-region replication, or CRR). We can do this for many use cases, like compliance and low-latency access.

• Notes about S3 Replication:

  • Versioning should be enabled in both source and destination buckets.
  • Buckets can be in two different accounts.
  • Replication of data is asynchronous.
  • The S3 service should have the correct IAM permissions to get data from the source and put data in the destination bucket.

• Deletions:

  • We can enable configurations to duplicate delete markers from the source to the destination bucket.

  • Permanent deletion of object versions is not replicated.

  • When configuring a bucket for replication, the existing data is not replicated to the destination bucket. To replicate existing data (or objects that failed replication), we must use S3 Batch Replication.

• **Replication isn’t transitive: This means that if bucket_1 is configured for replication to bucket_2, and bucket_2 is configured for replication to bucket_3, bucket_1 won’t be replicated in bucket_3.