Basic HyperSwap

Basic HyperSwap is another new availability feature that was announced on February 28, 2008. It was announced with the Preview of z/OS 1.10 but it will also be available on z/OS 1.9 (with APAR OA20658 - april 2008).

You may notice that HyperSwap isn't new and, of course, you are right. We already know the Full Function GDPS/PPRC with its HyperSwap functionalities. There's also a subset of this known as GDPS/PPRC/HyperSwap Manager. These HyperSwap functionalities basically automate the switch of all primary PPRC disk subsystems with the secondary PPRC disk subsystems in case of a planned or unplanned disk outage in one or more sites for z/OS, zVM and zLinux environments.

The "new Basic HyperSwap(TM) capability (to be enabled by TotalStorage(R) Productivity Center for Replication Basic Edition for System z), plans to provide a low-cost, single-site, high-availability disk solution which allows the configuration of disk-replication services using an intuitive GUI from z/OS. The intention is that with Basic HyperSwap function enabled, seamlessly swapping between primary and secondary disk volumes in the event of planned and unplanned outages such as hardware maintenance, testing, or device failure can be accomplished from z/OS". So unlike the earlier functions, Basic HyperSwap is a part of z/OS 1.9+, and it also has some limitations as opposed to the others :

• It is a single-site solution : source and target disk subsystems are on the same data center floor
• It only supports IBM storage (ESS 800, DS8000, DS6000).
• It only supports the z/OS environment

Basic HyperSwap in z/OS 1.9 will be managed by TPC-R (TotalStorage Productivity Center for Replication) software shipped with z/OS. TPC-R Basic Edition v3.4 will come as a no-cost feature of z/OS 1.9+.
How does it work ?
"IBM has done an analysis of all field storage subsystem failures and as a result has created a set of trigger events that are monitored by TPC-R. When one of these HyperSwap trigger events occurs, a ‘PPRC Data Freeze across all LSSs on all storage subsystems’ is invoked. All I/O to all devices is Queued (Extended Long Busy), thus maintaining full data integrity and cross volumes data consistency. z/OS then completes the HyperSwap operation of recovering the PPRC target devices and rebuilding all z/OS internal control blocks to point to the recovered PPRC target devices. When this has completed, all I/O is released and all applications continue to run against the recovered target devices, thus masking or replacing a complete disk subsystem outage, with a dynamic ‘busy’ and a redirection of all Host I/O".

If you want to find out more about Basic HyperSwap, its functionalities, prerequisites and the differences with GDPS/HyperSwap, there's a brand new z/OS Basic HyperSwap & GDPS HyperSwap Overview presentation on TechDocs. Be sure to read the presentor's notes as well as they contain a lot of additional information.