If an agent is running on the same box as a server, the partitioning algorithm should make that server the agent's primary, for the simple fact that the agent won't need to route its traffic outside of that machine's network adapter since both are co-located.
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-1024
mass add of key word FutureFeature to help track
Per triage, this has been moved to ON_QA with the expectation that is is fixed. If not, please reopen, otherwise close.
This was never fixed.
this can easily be "worked around" by using affinity - such that the agent has affinity with the server that is running on the same box.