If your environment includes multiple Platform Services Controller instances, and if one of the Platform Services Controller instances becomes unavailable, your environment continues to function. When the Platform Services Controller becomes available again, user data and other information are usually replicated within 60 seconds. In certain special circumstances, however, replication might take a long time.

Problem

In certain situations, for example, when your environment includes multiple Platform Services Controller instances in different locations, and you make significant changes while one Platform Services Controller is unavailable, you do not see replication across VMware Directory Service instances right away. For example, you do not see a new user that was added to the available Platform Services Controller instance in the other instance until replication is complete.

Cause

During normal operation, changes to a VMware Directory Service (vmdir) instance in one Platform Services Controller instance (node) show up in its direct replication partner within about 60 seconds. Depending on the replication topology, changes in one node might have to propagate through intermediate nodes before they arrive at each vmdir instance in each node. Information that is replicated includes user information, certificate information, license information for virtual machines that are created, cloned, or migrated with VMware vMotion, and more.

When the replication link is broken, for example, because of a network outage or because a node becomes unavailable, changes in the federation do not converge. After the unavailable node is restored, each node tries to catch up with all changes. Eventually, all vmdir instances converge to a consistent state but it might take a while to reach that consistent state if many changes occurred while one node was unavailable.

Solution

Your environment functions normally while replication happens. Do not attempt to solve the problem unless it persists for over an hour.