Asynchronous Replication

Asynchronous replication provides a time-lagged copy of data that is written to a secondary storage site, usually located remotely from the primary storage site.

Asynchronous replication requires at least two Replication Engines. Asynchronous replication is supported generally when the primary and secondary locations are geographically distributed, and communication is over a wide area network (WAN) link, with separate Replication Engines at each location. The default connectivity for the WAN connection is through the Ethernet management ports (eth0 and eth2). Whenever data is written to primary storage, a copy of that data is prepared for later transfer over a WAN connection to the secondary storage site.

For example, a typical asynchronous replication configuration consists of primary and secondary sites connected by an Internet protocol (IP) WAN. Each primary and secondary site consists of an Oracle MaxRep Replication Engine and an Oracle FS System.

When data from the application host is written to the primary Oracle FS System, a copy of the data, along with any necessary journal information, is prepared for transfer to the primary Replication Engine. The primary Replication Engine then transfers that data to the secondary Replication Engine, which writes the data to the secondary Oracle FS System.

Figure 1: Basic configuration for asynchronous replication

 Basic configuration for asynchronous replication 
Legend
1 Primary site 7 Source LUN
2 Secondary site 8 Secondary Oracle FS System
3 Host 9 Retention LUN
4 Replication Engine 10 Backup LUN
5 Home LUN 11 Target LUN
6 Primary Oracle FS System  
Note: If a service disruption on the primary site occurs, reverse replication is implemented. In this scenario, the target LUN becomes the source LUN, and the original source LUN becomes the target LUN. Reverse replication is activated when the original source site comes back online. In order for reverse replication to be effective, a retention LUN must be added to the original source location to facilitate the replication process.

Another example of asynchronous replication includes adding a secondary, or passive, Replication Engine to support the existing Replication Engine and to provide a high availability configuration. The primary Replication Engine and the secondary Replication Engine form an active‑passive cluster. The first Replication Engine in each clustered pair is in active mode. The second Replication Engine in each clustered pair is in passive mode, ready to take over if the active Replication Engine should fail.

A full high availability configuration consists of a cluster of two Replication Engines and the primary Oracle FS System on the primary site. The secondary site contains a cluster of two Replication Engines and the secondary Oracle FS System. High availability is not required at both sites. You can configure any Replication Engine as high availability by adding another Replication Engine.