Deployment Topologies with External Platform Services Controller Instances and High Availability

To ensure Platform Services Controller high availability in external deployments, you must install or deploy at least two joined Platform Services Controller instances in your vCenter Single Sign-On domain. When you use a third-party load balancer, you can ensure an automatic failover without downtime.

Platform Services Controller with a Load Balancer

Figure 1. Example of a Load Balanced Pair of Platform Services Controller Instances

You can use a third-party load balancer per site to configure Platform Services Controller high availability with automatic failover for this site. For information about the maximum number of Platform Services Controller instances behind a load balancer, see the Configuration Maximums documentation.

Important: To configure Platform Services Controller high availability behind a load balancer, the Platform Services Controller instances must be of the same operating system type. Mixed operating systems Platform Services Controller instances behind a load balancer are unsupported.

The vCenter Server instances are connected to the load balancer. When a Platform Services Controller instance stops responding, the load balancer automatically distributes the load among the other functional Platform Services Controller instances without downtime.

Platform Services Controller with Load Balancers Across vCenter Single Sign-On Sites

Figure 2. Example of Two Load Balanced Pairs of Platform Services Controller Instances Across Two Sites

Your vCenter Single Sign-On domain might span multiple sites. To ensure Platform Services Controller high availability with automatic failover throughout the domain, you must configure a separate load balancer in each site.

Platform Services Controller with No Load Balancer

Figure 3. Example of Two Joined Platform Services Controller Instances with No a Load Balancer

When you join two or more Platform Services Controller instances in the same site with no load balancer, you configure Platform Services Controller high availability with a manual failover for this site.

When a Platform Services Controller instance stops responding, you must manually fail over the vCenter Server instances that are registered to it. You fail over the instances by repointing them to other functional Platform Services Controller instances within the same site. See Repoint vCenter Server to Another External Platform Services Controller.

Note: If your vCenter Single Sign-On domain includes three or more Platform Services Controller instances, you can manually create a ring topology. A ring topology ensures Platform Services Controller reliability when one of the instances fails. To create a ring topology, run the /usr/lib/vmware-vmdir/bin/vdcrepadmin -f createagreement command against the first and last Platform Services Controller instance that you have deployed.

Platform Services Controller with No Load Balancer Across vCenter Single Sign-On Sites

Figure 4. Example of Two Joined Pairs of Platform Services Controller Instances Across Two Sites with No Load Balancer

Your vCenter Single Sign-On domain might span multiple sites. When no load balancer is available, you can manually repoint vCenter Server from a failed to a functional Platform Services Controller within the same site. See Repoint vCenter Server to Another External Platform Services Controller.

Important: Repointing vCenter Server between sites and domains is unsupported. If no functional Platform Services Controller instance is available in the site, you must deploy or install a new Platform Services Controller instance in this site. This new Platform Services Controller instance becomes the replication partner of the existing Platform Services Controller instance.