Experiences and tips from a team of system programmers and testers who run a Parallel Sysplex on which we perform the final verification of a z/OS release and System z hardware and System Storage before they become generally available to clients.
One requirement for installing an OpenShift cluster is a load balancer (LB) for directing traffic to the OCP nodes. Unfortunately, if the load balancer is unavailable, the cluster is not accessible. This experience report describes the process of configuring keepalived to share an IP address between two load balancers. This setup permits either load balancer to act as the primary load balancer with the other load balancer in standby, ready to take over within a few seconds should the primary go down. In this instance, Nginx is used for load balancing responsibilities.