Re: RequestLogger and session invalidation
On Thu, May 28, 2009 at 11:24 AM, Johan Compagner wrote: > why are you using invalidateNow? > I was using it for invalidating sessions on logout. But since plain invalidate() works, and doesn't break RequestLogger, my original problem is fixed. Regards, Taneli Korri
Re: RequestLogger and session invalidation
why are you using invalidateNow? On Thu, May 28, 2009 at 10:11, Taneli Korri wrote: > On Thu, May 28, 2009 at 6:49 AM, Jeremy Thomerson < > jer...@wickettraining.com > > wrote: > > > Please open a JIRA so this doesn't get lost. I haven't looked, but it > > sounds wrong if it truly recreates an invalidated session in some > > end-of-request logging, even if it doesn't bind the session. Post the > > link back here. > > > > > > I found out what's causing the problem, I was using Session.invalidateNow() > which invalidates the session before it's logged. With Session.invalidate() > the problem doesn't exist. > > > I created a JIRA issue about it: > > https://issues.apache.org/jira/browse/WICKET-2291 > > > Regards, > > Taneli Korri >
Re: RequestLogger and session invalidation
On Thu, May 28, 2009 at 6:49 AM, Jeremy Thomerson wrote: > Please open a JIRA so this doesn't get lost. I haven't looked, but it > sounds wrong if it truly recreates an invalidated session in some > end-of-request logging, even if it doesn't bind the session. Post the > link back here. > > I found out what's causing the problem, I was using Session.invalidateNow() which invalidates the session before it's logged. With Session.invalidate() the problem doesn't exist. I created a JIRA issue about it: https://issues.apache.org/jira/browse/WICKET-2291 Regards, Taneli Korri
Re: RequestLogger and session invalidation
Please open a JIRA so this doesn't get lost. I haven't looked, but it sounds wrong if it truly recreates an invalidated session in some end-of-request logging, even if it doesn't bind the session. Post the link back here. -- Jeremy Thomerson http://www.wickettraining.com On Wed, May 27, 2009 at 8:25 AM, Taneli Korri wrote: > Hi, > > I'm using Wicket 1.3.6 and RequestLogger.getLiveSessions to get the current > sessions in my web application, but I'm experience strange behaviour when > session invalidation occurs. > > When the user logs out of the application, the session is invalidated. This > leads to WebApplication.sessionDestroyed to be called, which removes the > session from the RequestLogger.liveSessions Map and everything seems to be > fine. > > But the execution continues and eventually RequestCycle.detach gets called, > which leads to RequestLogger.requestTime. And that's where the strangest > thing happens, on line 254 of RequestLogger.java the invalidated session is > created again and added back to the liveSessions Map. > > Is this the expected behaviour? Since this ensures that the liveSessions Map > includes all sessions, even the invalidated ones. > > > And another thing, are there any downsides in keeping RequestLogger on in > production environment? I'm thinking that it might hog a lot of resources, > but is that the case? > > > Regards, > > Taneli Korri > - To unsubscribe, e-mail: users-unsubscr...@wicket.apache.org For additional commands, e-mail: users-h...@wicket.apache.org
RequestLogger and session invalidation
Hi, I'm using Wicket 1.3.6 and RequestLogger.getLiveSessions to get the current sessions in my web application, but I'm experience strange behaviour when session invalidation occurs. When the user logs out of the application, the session is invalidated. This leads to WebApplication.sessionDestroyed to be called, which removes the session from the RequestLogger.liveSessions Map and everything seems to be fine. But the execution continues and eventually RequestCycle.detach gets called, which leads to RequestLogger.requestTime. And that's where the strangest thing happens, on line 254 of RequestLogger.java the invalidated session is created again and added back to the liveSessions Map. Is this the expected behaviour? Since this ensures that the liveSessions Map includes all sessions, even the invalidated ones. And another thing, are there any downsides in keeping RequestLogger on in production environment? I'm thinking that it might hog a lot of resources, but is that the case? Regards, Taneli Korri