Affects Version/s: 4.2.0
Fix Version/s: None
Component/s: Core Monitoring
Recently a log line was added to detect metrics coming from the wrong agent:
2010-02-03 14:52:27,571 WARN [Thread-1724] [org.hyperic.hq.measurement.server.session.ReportProcessorEJBImpl@210] measurement (id=14813) was sent to the HQ server from agent (agentToken=1265054785885-5955226501439245874-8190015691460643997) but resource (id=11018) is not associated with that agent. Dropping measurement.
In the case where a network device is reconfigured to use a different agent, the metrics scheduled on the original agent are not unscheduled causing this message to fill in the logs.
Steps to reproduce:
Agent 1 - IP 10.0.0.1
Agent 2 - IP 10.0.0.2
In this example, agent 1 is monitoring a network device at 10.0.0.111. If the user navigates to that devices inventory page and changes the agent connection from 10.0.0.1:2144 to 10.0.0.2:2144 we correctly reschedule the metrics on the new agent, but the old agent at 10.0.0.1 continues to collect these measurements. This results in the logs filling with the error noted above.