Synchronous Replication of DB2 UDB for z/OS

Use

You can use synchronous replication of DB2 database data if your site:

·        Requires the standby system to always be fully up-to-date with the primary system

·        Can accept some performance impact on write operations at the primary location

You can use Geographically Dispersed Parallel Sysplex (GDPS) to copy data synchronously to the standby site.

Features

Geographically Dispersed Parallel Sysplex (GDPS), also called "Geoplex", is a Parallel Sysplex cluster spread across two sites, with all critical data mirrored between the sites using Peer-to-Peer Remote Copy (PPRC). It offers a controlled site switch for both planned and unplanned site outages, with no data loss, maintaining full data integrity in the SAP system.

PPRC is a high availability solution that provides storage-based disaster recovery and workload migration with the capability to copy data in real time to a remote location.

A PPRC data copy to the standby system is simultaneously synchronized with the I/O operation on the primary system. This means that PPRC does not consider the primary system write operation complete until it has received a signal from the standby system that the operation is complete. This signal is sent from the disk system of the standby system after the data is saved in the cache. The response delay increases the primary system's response time.

Geoplex consists of a Parallel Sysplex cluster spread across two sites with one or more z/OS systems at each site. The maximum distance between the two sites is 40 kilometers.

Geoplex supports the following configuration options:

·        Single site workload

This configuration is intended for those systems that have critical workload on the primary site and expendable work on the standby site.

·        Multiple site workload

This configuration is intended for those systems that have critical and expendable workload on both sites.

In this section only the single site workload configuration is described. For more information about the multiple site workload see the IBM documentation Geographically Dispersed Parallel Sysplex: the S/390 Multi-site Application Availability Solution.

Geoplex consists of the following systems:

·        Primary systems

The primary systems execute mission-critical work.

·        Standby systems

The standby systems do not run any mission-critical work. They can run work that is considered expendable. If one or more of the primary systems is unavailable, expendable work on the standby systems is slowed down or stopped and processing resources are provided to take over the mission-critical work.

·        Controlling system

The controlling system coordinates switchover.

Geoplex uses Parallel Sysplex cluster facilities to communicate between systems. For example, each Geoplex system joins the Geoplex Parallel Sysplex cluster group. Each system monitors the Parallel Sysplex cluster, coupling facilities, and storage subsystems and maintains Geoplex status. The following graphic illustrates a possible configuration and is used for the description in the section "Activities" below:

Geoplex with Parallel Sysplex

Activities

You configure Geoplex so that:

·        The mission-critical SAP system is running on the primary systems SA1 and SA2, located in the primary site A. In this example DB2 data sharing is used but you can also run the database server without DB2 data sharing in the Geoplex environment.

·        Two standby system on site B provide processing resources when one or more of the  primary systems are unavailable or site A is unavailable.

·        The controlling system SB1 monitors the Geoplex configuration and coordinates Geoplex processing.

·        Database DA on primary site A is mirrored to the standby copy on site B.

The following graphic shows the situation after site A fails or is shut down.

Geoplex Following Failure or Shutdown at Site A

After shutdown or failure at site A:

...

       1.      Geoplex freezes the standby copy of data on site B to guarantee data consistency. PPRC has stopped working because site A is not operational.

       2.      Geoplex initiates a site takeover so that the:

Ў        Standby copy of the data is now considered the primary copy

Ў        Active standby system SB1 expands

Ў        Inactive standby system SB2 is started to acquire processing resources needed to run the SAP system

Ў        Database server is started on the systems SB1 and SB2 on site B

       3.      The SAP system administrator starts the application servers on the standby site.

When site A resumes operation:

·        z/OS operator can initiate the switch from site B to site A.

·        Geoplex restores the original configuration.

·        The SAP system administrator starts the application servers on the primary site.

After shutdown or failure at site B:

·        Remote copy processing is suspended.

·        One of the primary systems in site A assumes the role of the controlling system.

·        The SAP system remains available.

When site B resumes operation:

·        z/OS operator can initiate the switch from site A to site B for the controlling system.

·        Geoplex restores the original configuration.