Hi,
I saw the log book entries and emails about the monitoring app being a prolific "thread factory" that spawned many timer threads (which were used to redraw the plots), and this eventually locked up the machines in the counting house when the app was left on for a long time.
The problem is tracked by this JIRA issue.
https://jira.slac.stanford.edu/browse/HPSJAVA-504
So I believe that I have fixed this by rewriting some code in the monitoring app plotting backend which does the plot updates. This is deployed to the Maven repo as jfreechart-aida-experimental 1.7-SNAPSHOT (prior version we were using wa 1.6). I have also updated HPS Java trunk to use this new version.
My testing on EcalMonitoringFinal.lcsim with run 5077 indicates this is *much* better in terms of thread management. There should only be a few threads now that are used for redrawing the plots, and the timer threads no longer should pile up as the app runs. Please test this new configuration and let me know if it fixes the issue.
Thanks.
--Jeremy
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
|