Federator does not remove obsolete topology objects when children are in HA (4375274)

Federator does not remove obsolete topology objects when children are in HA

When a federator master has connected children in HA mode and objects are deleted from a child server, the objects are not deleted from the federator master. This requires the user to manually delete objects both on the child Management Servers as well as the federator.

The following messages will be present in the Foglight Management Server logs:
WARN    [RemoteTopologySync-10-thread-17] com.quest.nitro.service.federation.TopologySynchService - Obsolete federated topology object purge will not be performed because the following federated servers are not connected: [rmi://fms01:1099, rmi://fms02:1099].

When the Foglight Management Server has connected children in HA mode, the secondary servers are considered as disconnected. As a result, the obsolete federated topology object purge operation does not execute.

A system property is available to ignore disconnected child servers and force object purge on the federation master (default is false).

To enable, run the following script from the Foglight Script Console (Running a groovy script using the script console (4309796)):

System.setProperty("foglight.federation.force_purge", "true")

Note: Once enabled, wait for a new "Major refresh start" message in the Management Server logs before validating if objects are being removed.

It is recommended to use this option only for cleanup purposes rather than enabling it permanently. Otherwise, whenever a child server loses connectivity, the subsequent topology synchronization will initiate an unintended purge. To disable it, set the property value back to false:

System.setProperty("foglight.federation.force_purge", "false")

Note: The property value will also revert back to the default false value if the Federator is restarted.

FGL-20233