[undertow-dev] sessionId changes between requests?

Stuart Douglas sdouglas at redhat.com
Fri Jan 22 00:14:41 EST 2016


Something to be aware of is that in Servlet 3.1 users can also trigger this change by calling javax.servlet.http.HttpServletRequest.changeSessionId(). 

Not sure if that will also cause issues for you or not.

Stuart

----- Original Message -----
> From: "Bill Burke" <bburke at redhat.com>
> To: "Stuart Douglas" <sdouglas at redhat.com>
> Cc: undertow-dev at lists.jboss.org
> Sent: Friday, 22 January, 2016 3:29:29 PM
> Subject: Re: [undertow-dev] sessionId changes between requests?
> 
> Maybe a decoupling of cookie from session ID isn't very feasible...I
> guess I can just turn off the "changeSessionIdOnLogin" switch and change
> the ID within the authenticator instead.
> 
> On 1/21/2016 10:28 PM, Stuart Douglas wrote:
> > This was done for security reasons (see
> > https://issues.jboss.org/browse/UNDERTOW-579).
> >
> > I don't know how practical it would be to de-couple the cookie value from
> > the session ID. Could you just use a
> > javax.servlet.http.HttpSessionIdListener to monitor session ID changes?
> >
> > Stuart
> >
> > ----- Original Message -----
> >> From: "Bill Burke" <bburke at redhat.com>
> >> To: undertow-dev at lists.jboss.org
> >> Sent: Friday, 22 January, 2016 11:44:53 AM
> >> Subject: Re: [undertow-dev] sessionId changes between requests?
> >>
> >> Ok, found it.  setChangeSessionIdOnLogin()
> >>
> >> Can I ask why this is done?  Security reasons?  To change the cookie?
> >> If it is to change the cookie, would be really good in the future to
> >> decouple the session cookie value from the session id so that plugins,
> >> like Keycloak, that are remotely managing and monitoring sessions can
> >> still do so without creating a security hole.
> >>
> >> On 1/21/2016 6:10 PM, Bill Burke wrote:
> >>> Does a HttpSession ID change between requests?  We are storing the
> >>> current HttpSession ID at our IDP after login, then transmitting back to
> >>> the app in a background HTTP request, looking up the session and then
> >>> invalidating it. This used to work on Wildfly 8 and 9, in 10, looks like
> >>> it is not the same http session.
> >>>
> >> --
> >> Bill Burke
> >> JBoss, a division of Red Hat
> >> http://bill.burkecentral.com
> >>
> >> _______________________________________________
> >> undertow-dev mailing list
> >> undertow-dev at lists.jboss.org
> >> https://lists.jboss.org/mailman/listinfo/undertow-dev
> >>
> 
> --
> Bill Burke
> JBoss, a division of Red Hat
> http://bill.burkecentral.com
> 
> 


More information about the undertow-dev mailing list