* primary server re-computes agent distribution when: ** group membership events occur ** agent load changes (more than some threshold) * distribution lists are sent down to each server and persisted to the database
This actually encompassed a lot of stuff - The base distribution algorithm and unit tests - Persistence handling of partition events and server (failover) lists - Startup/Shutdown/HA change hooks for all ParitionEventTypes for audit, immediate or requested partition execution - Agent services for getting the server lists - single quartz job ClusterManagerJob for managing cluster events - server instance quartz job ServerManagerJob for managing server events and keeping a server heartbeat - various schema changes
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-674