Replication Workflow

Overview

The replication workflow feature in ARTESCA enables users to replicate data from a source bucket to one or multiple target cloud storage locations.

This feature helps in achieving higher data durability and accessibility by maintaining copies of data in multiple locations. It also ensures data redundancy, availability, and disaster recovery.

Architecture

	Cross-Region Replication Location	Other Locations
Active-Passive [1]	Supported	Supported
Active-Active [2]	Supported	Not Supported

[1]

Active-Passive architecture is achieved through a uni-directional replication workflow where data flows from a source site to a destination site.

[2]

Active-Active architecture is achieved through a bi-directional replication workflow where data replicates in both directions between the two sites, allowing updates on either site to be synchronized to the other.

Behavior

	Cross-Region Replication Location	Other Locations
Asynchronous Delay	Replication is asynchronous and event-driven, meaning objects appear on the destination with a delay after being written to the source site.
Object Version Order	Identical between source and destination.	Not preserved
Object Deletes (soft deletes)	Replicated	Replicated
Object Version Deletes (hard deletes)	Not replicated	Not replicated
Object Version IDs	Identical between source and destination.	Not preserved
Object Retention Lock	Identical between source and destination.	Not replicated
Object Tags/Metadata	Identical between source and destination.	Identical between source and destination.
Conflict Rule	In bi-directional setups, conflicts are resolved using a last-writer-wins rule.	Object Version Order is not preserved.
Encryption	Source object encryption settings are preserved on the destination.	-–
Access Control Lists (ACLs)	Object ACLs are replicated, but cross-account permissions can require additional configuration on the destination site.	Not replicated

Limitations

	Cross-Region Replication Location	Other Locations
Number of Destinations Per Configuration	Each replication configuration can only target a single CRR location.	Data from a source bucket can be replicated to one or multiple target cloud storage locations.
Maximum Number of Destination Buckets for Replication	Unlimited The destination bucket is specified in the replication rule. A single CRR location can be used to configure replication to different buckets on the remote site. A maximum of 10 locations can be used.	10 The destination bucket is specified in the storage location definition. One location is configured for each remote bucket (on target site) which will be used as a replication target. A maximum of 10 locations can be used.
Chaining [3]	Not Supported	Supported

[3]

Chaining refers to the replication across three or more sites (e.g., A -> B -> C).

Requirements

	Cross-Region Replication Location	Other Locations
Required Buckets Configuration	Source and destination buckets must have versioning enabled.
Required Roles	A source role (arn:aws:iam::root:role/s3-replication-role) and a destination IAM role.	-–
Required Destination Policies	Bucket policies on the destination must explicitly allow the s3:ReplicationObject action or the replication will be denied.	-–

Configuration

Refer to the following procedure to configure the replication workflow:

• Create a Replication Workflow

Operation

Refer to the following procedures to operate the replication workflow:

• Update a Replication Workflow
• Delete a Replication Workflow
• Configure the Replication Retry Mechanism

Tip

To tune the replication workflow, contact Scality support and refer to Scaling Replication Services.

Monitoring

Refer to the following page for more information on monitoring the replication workflow:

• Replication Monitoring