> OK, so is this a WAR installation that you just dropped onto Tomcat? Just
> want to be totally sure I'm using the same setup.

Yes, exactly.

> I'll set up a test just to see if I experience the same thing and if so
> we'll hand it off to Alan and Andy for verification and a fix if necessary.

I don't know exactly how this works "under the hood", but it sounds
almost like a seperate process which every X minutes take care of
cleaning the timed-out sessions. And a missing check if a session is
used again to see if it hasn't already timed out.

> Appreciate the thorough information--we'll get to the bottom of it!

Well, I remembered a very good presentation from Ben Nadel on the
"power" of just Application.cfc. Also I remembered several cases with
big problems around sessions/cookies (only recently, with the session
fixation "fix" in CF9.0.1 hotfix2) and that is why I was playing
around with this.

Which leaves the "client variables". Personally I like to use them
because I don't like the session-to-server binding in a load-balanced
environment. So if apps need to scale, I prefer to make them
sessionless. I understand the limitations of client variables over
sessions (only basic structures), but the session-in-database aspect I
like. But please tell me your opinion why to avoid using clients and
how do we solve that on load-balanced environments?

What also comes out of this "testing" is that I now understand the
setclientcookies better. If I am not mistaken, I could disable
sessions and clients and still have

<cfset this.setclientcookies = "Yes">

... assign unique CFID&CFTOKEN cookies for visitors. Maybe build
alternative client management with that?

Thanks again for looking into this.

Ivo

-- 
online documentation: http://openbd.org/manual/
   google+ hints/tips: https://plus.google.com/115990347459711259462
     http://groups.google.com/group/openbd?hl=en

     Join us @ http://www.OpenCFsummit.org/ Dallas, Feb 2012

Reply via email to