Thanks for the explanation, Marvin! You said that you haven't used Spring Webflow much. So this question is for people who know more about it. Would Spring Webflow store more information in the session the longer the session lasts?
> Simply that increasing session duration by a factor of 48 (240/5) > would cause a proportional increase in memory usage because the > session data would not be liable to garbage collection until the > session timed out. You are right! Session data would not be reclaimed by the garbage collector until the session expires. I did not think about it from that angle! Thanks for pointing that out. However, if an application doesn't store much data in the session, then it probably doesn't matter much. Hence its important to know what CAS stores in an HTTP Session. > Does CAS run on this machine as well? In any case I understand that > you have only a single CAS instance, which is really more germane to > the issue. Are you certain you have only a single Quartz job > definition? I'd recommend grepping for multiple job definitions if > you have not; the following might be a good search: > Yes, CAS runs on the same machine. Yes, only one CAS webapp. Yes single Quartz job. I tried to analyze the cas.log file and here's what I found: 1) The last entry was on Oct 2, Friday! I wonder why the log file isn't getting updated anymore? 2) The DefaultTicketRegistryCleaner starts every 83 minutes. But it runs multiple times before it stops and runs again after 83 minutes. To see what I am talking about, please visit: http://pastebin.com/m62c797d6 Thank you, Joe On Wed, Oct 7, 2009 at 11:40 AM, Marvin Addison <[email protected]> wrote: >> What are the objects that are stored in a CAS session? > > CAS uses Spring Webflow to manage the login workflow process. By > default all state management for the workflow is stored in session > data. Although I've done neither measurement nor research into the > memory footprint of Spring Webflow, I would imagine it could be > relatively large. > >> Can you please >> explain the cause-effect relationship you are talking about? > > Simply that increasing session duration by a factor of 48 (240/5) > would cause a proportional increase in memory usage because the > session data would not be liable to garbage collection until the > session timed out. > >> Our Apache/Tomcat server (running on a RedHat Enterprise Linux ES 4.0 >> machine) hosts 3 websites. CAS provides Single Sign-On service to >> these 3 websites. > > Does CAS run on this machine as well? In any case I understand that > you have only a single CAS instance, which is really more germane to > the issue. Are you certain you have only a single Quartz job > definition? I'd recommend grepping for multiple job definitions if > you have not; the following might be a good search: > > grep -R DefaultTicketRegistryCleaner $CATALINA_HOME/webapps/cas/* > > M > > -- > You are currently subscribed to [email protected] as: > [email protected] > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user > -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-user
