On Tue, Jan 25, 2011 at 1:35 PM, Adib <amsl...@gmail.com> wrote: > Thanks for the link that was useful. So the key idea to understand is > that there is a background thread that does all the cleanup and this > background thread executes on a timer not on very request arriving a > the server. > > I am surprised at the limitations of the Persistent Manager it's that > it should be an easy thing to backup and load sessions on every > request. Is there something about how tomcat is implemented that makes > implementing persistence on load and save difficult? > Not really. I'd assume that there were specific requirements that lead to this solution (how PersistentManager is implemented). Saving/loading sessions per request from/to a datastore can be done using a valve.
I just just implemented what you're looking for with the http://code.google.com/p/memcached-session-manager/ to support non-sticky sessions (to be released in a few days). I chose memcached as backend to be really scalable, as IMHO with a database this is expensive / hard to achieve. AFAIK Reinwald is just checking out do achieve what you want with PersistentManager: http://old.nabble.com/Why-cant-the-the-classes-%28in-the-jars%29-places-in-Tomcat-lib--see-the-classes-from-the-webapp-WEB-INF-lib.-td30713002.html#a30720080 Cheers, Martin > > Cheers > Adib > > On Tue, Jan 25, 2011 at 4:24 AM, Reinwald Warapen > <reinwal...@directi.com> wrote: > > On 1/25/2011 5:02 PM, Adib wrote: > >> > >> Hi, > >> > >> I am trying to understand the settings on the persistent session > >> manager on tomcat 6 and 7. It seems that the persistent session > >> manager is primarily meant for the purpose of swapping IDLE sessions > >> to persistent storage and then hope that they expire, I can see the > >> value of this given that many users abandon sites without logging out. > >> It also seems that the persisent manager is not designed to persist > >> every session at the end of every request so that at the start of > >> every request the session is read from the store and the end of the > >> request the session is written to store, is this understanding > >> correct. > > > > > http://reinwaldwarapen.com/2011/01/17/storing-and-sharing-sessions-among-standalone-tomcat-instances/ > >> > >> I have also been puzzling over the meanings of maxIdleSwap vs. > >> maxIdleBackup not sure why these two settings exist or what is > >> difference between them and the practical use case for them? It is > >> clear to me what the idle session is but what is meant by backup and > >> how is it different than a swap? > >> > >> Thanks > >> Adib > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > >> For additional commands, e-mail: users-h...@tomcat.apache.org > >> > >> > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > > For additional commands, e-mail: users-h...@tomcat.apache.org > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > For additional commands, e-mail: users-h...@tomcat.apache.org > > -- Martin Grotzke http://www.javakaffee.de/blog/
