What is Replication?

Replication maintains a copy or replica of a volume or set of volumes and their snapshots on another array that is configured as one of a pair of replication partners. The replica contains the contents of a volume at the time the replica was created or last updated, as well as a configured number of prior states (snapshots). Replicas are stored at a remote array, called a replication partner, connected by a network or Internet link. A volume is always located on a different array than its replica. It is possible to retain more or fewer snapshots on the replica than on the source volume, thus providing greater flexibility in designing a recovery plan.

The HPE Peer Persistence feature is designed for applications that require zero to near zero RPO and RTO storage. The feature enables multi-site synchronous replication with automatic switchover (ASO), which allows your arrays to recover automatically and non-disruptively from a storage based failure. For the purposes of this documentation, the two components of HPE Peer Persistence are referred to by their functional descriptions, synchronous replication and ASO, rather than the feature name.

NOTE: The HPE Peer Persistence feature is designed to enable applications to achieve a near zero RPO only in the event of storage failures. The feature will not prevent site-to-site failures that result from host level failures.

A replica is a copy of a volume from another group (in the case of snapshot replication) or pool (in the case of synchronous replication) whose state is managed by the other group or pool. For example, all write operations to a replica originate from another group. The array that hosts the original volume and manages the state of a replica is called the upstream partner, because the data flows from it. The array where the replica resides is called the downstream partner, because the data flows to it.

The replicated volumes can be restored as complete copies of the volumes, with all schedules and administrative settings replicated, as well as the actual data. Like snapshots, replicas are created and stored based on volume collection schedules. Multiple volumes can share a volume collection schedule.

For snapshot replication only: When you promote a replica, the number of snapshots to retain is adjusted to be the maximum number of snapshots to retain on the local array, plus the number of snapshots to retain on the replica. The volume is offline until a replica is promoted. The volume and settings are visible on the replica, but they are not editable until a volume collection handover is performed.

NOTE: For synchronous replication, the snapshot retention limits remain with each pool after a handover.