Monitord "hangs" with mysql backend

Recently, I have seen monitords staying around after workflows finishes. I'm not sure if this is a new issue or not.

I gave monitord 6 SIGUSR1s to increase the debug level. After a few minutes I got:


2011-07-01 18:11:19,058:nllog.py:165: DEBUG:
ts=2011-07-01T22:11:19.058352Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.reset_flush_state
level=Debug msg="Resetting flush state"
2011-07-01 18:11:19,058:nllog.py:165: DEBUG:
ts=2011-07-01T22:11:19.058737Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.hard_flush.end
level=Debug batching=1
2011-07-01 18:11:19,059:nllog.py:165: DEBUG:
ts=2011-07-01T22:11:19.059003Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.reset_flush_state
level=Debug msg="Flush: flush count"
2011-07-01 18:11:19,059:nllog.py:165: DEBUG:
ts=2011-07-01T22:11:19.059391Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.process
level=Debug msg="{'child.task.id': 'ID006163', 'parent.task.id':
'ID001117', 'event': 'stampede.task.edge', 'xwf.id':
'04139283-fdec-4601-ac2d-3a7af07a4553', 'ts': 1309539279}"




But then, it kind of hangs again:

2011-07-01 18:11:24,930:nllog.py:165: DEBUG:
ts=2011-07-01T22:11:24.930016Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.hard_flush.begin
level=Debug batching=1



And then 10 minutes later:

2011-07-01 18:22:47,488:nllog.py:165: DEBUG:
ts=2011-07-01T22:22:47.487925Z
event=netlogger.analysis.modules.stampede_loader.Analyzer.reset_flush_state
level=Debug msg="Resetting flush state"