Dell PowerVault ME5 Series Administrator's Guide

Replication prerequisites

To replicate a volume, you must first create a peer connection or use an existing one, and create a replication set. A peer connection establishes bi-directional communication between a local and remote system, both of which must have FC or iSCSI ports and a virtual pool. The system establishes a peer connection by connecting a host port on the local system with a user- specified host port on the remote system, then exchanging information and setting up a long-term communication path in- band. Because the communication path establishes a peer connection between the two systems, replications can occur in either direction.

To verify that a host port address is available before creating a peer connection in the PowerVault Manager, specify a peer system IP address and then choose Query Peer Connection (Settings > Peer Connections). Alternatively, use the query peer-connection CLI command. This command provides information about the remote system, such as inter-connectivity between the two systems and pool configuration. For more information on this command, see the CLI Reference Guide.

In the PowerVault Manager, creating a peer connection or selecting an existing one, is part of creating a replication set. After you create or select a peer connection, you can continue to create a replication set. A replication set specifies one or more volumes, multiple volumes in a volume group, or snapshots on one system of the peer connection, known as the primary system in the context of replication, to replicate across the peer connection. When you create a replication set, corresponding volumes are automatically created on the other system of the peer connection, known as the secondary system, along with the infrastructure needed for replication. The infrastructure consists of internal snapshots used for replication operations. A replication set for a volume consumes two internal snapshots each for the primary volume and the secondary volume if the queue policy is set to Discard, or three each if the queue policy is set to Queue Latest. A replication set for a volume group consumes two internal volume groups if the queue policy is set to Discard, or three if the queue policy is set to Queue Latest. Each internal volume group contains a number of volumes equal to the number of volumes in the base volume group.

Using a volume group for a replication set enables you to make sure that multiple volumes are synchronized at the same time. When a volume group is replicated, snapshots of all of the volumes are created simultaneously. In doing so, the volume group functions as a consistency group, ensuring consistent copies of a group of volumes. The snapshots are then replicated as a group. Even though the snapshots may differ in size, replication is not complete until all of the snapshots are replicated.

For a replication set, the term primary refers to the source volume and the system in which it resides, and the term secondary is used for the remote copy and the system in which it resides. The secondary volume is meant to be an exact copy of the primary volume from the last time that replication occurred. To guarantee that the contents from that point in time match, the secondary volume cannot be mapped, rolled back, or modified except through replication.

While you cannot modify the secondary volume, you can create a snapshot of the secondary volume that you can map, mount, roll back, and otherwise treat like any volume or snapshot. You can regularly take snapshots to maintain a history of the replications for backup or archiving, or enable snapshot history for the replication set. These snapshots also can be used in disaster recovery.