Combining Synchronous and Asynchronous Replication

Note

This feature is introduced in VAST Cluster 5.4.1 and requires all participating clusters to be running VAST Cluster 5.4.1.

Overview

VAST Cluster supports a combination of synchronous and asynchronous replication where three clusters participate in the configuration. This option enables you to run S3 synchronous replication to a secondary cluster for resiliency with zero data loss in the event of disaster affecting only the primary cluster. It also provides an additional backup for both sync peers on a third async peer cluster.

In the event of the primary cluster becoming inaccessible, recovery is achieved by turning over to the secondary cluster. As an additional backup in the event that both the primary and secondary sync replication peers are unavailable, recovery is achieved by failing over to the async replication peer. You can then fail back again to the primary cluster in the event that it becomes available.

To deploy this type of replication, you configure:

• A protected path with on a source path on one cluster (cluster A) with sync replication to a destination path on a second cluster (cluster B).

• An additional destination (cluster C) on the same protected path, with asynchronous replication.

• Standby connections between clusters B and C. These connections enable disaster recovery scenarios.

How to Configure Combined Sync and Async Replication

1. Make sure there is a virtual IP pool on each of the peer clusters that is dedicated to replication. The role of the virtual IP pool must be set to replication.
   A replication VIP pool is used exclusively for routing replication traffic between the peers and not for serving data to clients. The CNodes that are assigned virtual IPs from the replication virtual IP pool are used to communicate directly with the remote peer, while other CNodes can communicate only indirectly with the remote peer.

   If there is no such virtual IP pool on any of the clusters, follow the procedure to create a virtual IP pool, and be sure to set the role to Replication.

2. If you want to configure replication in secure mode, with mTLS encryption, make sure that mTLS certificates are installed on every participating cluster.

3. On the cluster A (the primary cluster):

   1. Create a replication peer.  for replication from A to B and from A to C. Each of these replication peer configurations is mirrored on the respective destination cluster.

   2. Create a protection policy  for asynchronous replication to cluster C (the asynchronous replication peer). This policy is automatically mirrored on cluster C.

   3. Create a remote protected path. Configure a new (previously non existent) path on cluster B as the destination with synchronous replication.

   4. Modify the protected path to add another path on cluster C as a destination with asynchronous replication.

4. On cluster C:

   1. Create a replication peer to connect cluster C to cluster B.

   2. Create a protection policy for replication from C to B.

   3. Identify the protected path that you created on cluster A and update it. Add cluster B as a standby destination.

Disaster Recovery Scenarios

See Recovery Scenarios for Combined Sync and Async Replication.