Description of problem: After stopping machine with sub collector the daemons are still present in master collector statistics. Version-Release number of selected component (if applicable): condor-7.6.5-0.14 How reproducible: 100% Steps to Reproduce: 1. Set sub collector on remote machine COLLECTOR0: $(COLLECTOR) COLLECTOR0_ARGS = -f -p 12345 COLLECTOR0_ENVIRONMENT = "_CONDOR_COLLECTOR_LOG=$(LOG)/CollectorLog" DAEMON_LIST = MASTER, STARTD, COLLECTOR0 CONDOR_HOST = <master.collector.machine> STARTD.COLLECTOR_HOST = <sub.collector.machine>:12345 CONDOR_VIEW_HOST = $(COLLECTOR_HOST) 2. check running daemons using 'condor_status -any' # condor_status -any MyType TargetType Name Collector None Personal Condor at <sub.collector.machine> Machine Job <sub.collector.machine> Machine Job <sub.collector.machine> DaemonMaster None <sub.collector.machine> Negotiator None <master.collector.machine> Collector None Personal Condor at <master.collector.machine> ... 3. stop condor on machine with sub collector, check the daemons again # condor_status -any MyType TargetType Name Collector None Personal Condor at <sub.collector.machine> Machine Job <sub.collector.machine> Machine Job <sub.collector.machine> Negotiator None <master.collector.machine> Collector None Personal Condor at <master.collector.machine> ... Actual results: Only ads from directly connected daemons were removed, DaemonMaster in this case. Sub collector and daemons connected to it are still present. They are removed after 15 minutes. Expected results: All daemons should be removed immediately after stopping condor on machine with sub collector. Additional info:
FYI, Version should likely be 1.0. This is by design. Consider, TOP-Collector <- { SUB-Collector-A, SUB-Collector-B } <- startd. Shutdown of SUB-Collector-A should not erase knowledge of startd from TOP-Collector.