Synchronous PPRC, also known as Metro Mirror, establishes a relationship between a primary source volume and a secondary copy volume in which both are updated simultaneously. The synchronous copy ensures that I/O completion of the application write to the primary is not signalled until after the write to the secondary is also complete. In a synchronous PPRC relationship, the secondary is always identical to the primary.
The advantage of a synchronous PPRC solution is that there is minimal host impact for performing the copy, since it is all handled by storage control hardware. No host resources are used to perform the initial copy or shadow subsequent updates. This is attractive to installations with limited CPU resources. The secondary storage control does not necessarily need to be channel attached to the application host, just the recovery host. This is attractive to installations with limited application host addressability.
The disadvantage is that since the copy operation is synchronous, there is an impact to application performance, as the application I/O operation will not be signalled as complete until the write to the secondary is also complete. The longer the distance between primary and secondary storage controllers, the greater this impact to application I/O, and therefore, application performance.
Plan your configuration to provide for capacity requirements, redundancy requirements, and performance requirements. Peer-to-peer remote copy allows from one to eight ESCON® or FCP paths (IBM® recommends at least two paths) from any primary site logical subsystem to a single recovery site logical subsystem. A specific primary site logical subsystem can connect with up to 16 different recovery site logical subsystems. However, you can link a single logical subsystem at the recovery site to a maximum of 64 primary site logical subsystems. A large cache and NVS will help the recovery site logical subsystem to accept the copy work loads from multiple primary site logical subsystems.
When you are establishing paths, distribute them between different storage clusters (ESS) and Directors to gain flexibility and availability. If you have a path that involves an LSS to LSS pairing you can only have one type of path, ESCON or FCP. Paths for a source or target LSS can be mixed, as long as the different paths do not involve the same source and target LSSs.
ESCON | FCP |
---|---|
Primary Storage Subsystem
|
Primary Storage Subsystem
|
LINK
|
LINK
|
Secondary (Recovery) Storage Subsystem
|
Secondary (Recovery) Storage Subsystem
|
For additional information about the CESTPATH command, with its syntax and parameters, refer to CESTPATH – establishing paths.
For additional information and examples about establishing PPRC paths, refer to Establishing PPRC paths.