High availability of IBM z/OS Connect Servers

If you have multiple z/OS Connect Servers configured, you can implement a high availability solution to provide scalability and backup in case of a server failure

Started task Applies to z/OS Connect Servers run by using a z/OS started task procedure.

HA can be implemented with just two z/OS Connect Servers that run in a single LPAR. Or you can design a more complex parallel sysplex solution with many z/OS Connect Servers that are spread across multiple LPARs.

The following diagram illustrates how z/OS Connect works within an HA environment across multiple LPARs.

Figure 1. Topological characteristics of an HA configuration
Illustration of the HA environment for the API requesters scenario.

zosConnect-2.0 Each z/OS Connect Server should have a unique HTTP or HTTPS port that is dedicated for use by the RESTful administration interface to allow administration requests to be directed to the appropriate server. In the Figure 1, these are shown as ports 1001, 1002, 1003 and 1004. For more information, see Management of IBM z/OS Connect APIs, services, and API requesters in an HA environment.

High availability can be achieved with z/OS Connect by the workload distribution of HTTP and HTTPS API and requests from clients across several servers with shared configuration. For more information, see Connectivity for high availability and Sharing server configuration. To ensure that workload distribution can route requests to any configured z/OS Connect Server, the servers must be maintained as clones of each other. That is, the servers must have the same configuration with the same z/OS Connect APIs, or API requesters deployed.

For a step-by-step example of how to implement high availability for inbound requests to two z/OS Connect Servers that use port sharing and shared configuration data, see Configure HA for connections to IBM z/OS Connect.