Lost session

[Xiaoshu Wang]

Thanks Scott. I guess it might have something to do with Glassfish as well. I wish I can use the environment variables. But the group who managed that is not willing to do that because of the ease of administration and I don't have much say in that.

Thanks for the help. At least I know what I can and I cannot do.  

-- 
Xiaoshu Wang
Sent with Sparrow (http://www.sparrowmailapp.com/?sig)


On Tuesday, September 3, 2013 at 9:58 PM, Cantor, Scott wrote:

> On 9/3/13 9:39 PM, "Xiaoshu Wang" <xiaoshuw at email.unc.edu (mailto:xiaoshuw at email.unc.edu)> wrote:
> 
> > Just did a test, hope this might help shine some lights on.
> 
> I know nothing about Glassfish, I'm just telling you your basic
> interpretation is off in some way.
> 
> > I put in these configuration on glassfish-web.xml (http://web.xml) because without it.
> > When sending multiple request, some request is returned back (200 code
> > but with an empty response body). My server code never see these requests.
> > 
> 
> 
> If the SP were trying to cause a session to be created, as you were
> suggesting, those would be 302 redirects, not empty responses.
> 
> > And all these happens when there is
> > a valid shibboleth session. However, turning off these glassfish option,
> > then the previous problem is gone.
> > 
> 
> 
> I guess the glassfish code is doing something then, but that isn't
> anything I can speak to.
> 
> > <session-config>
> > <session-manager persistence-type="replicated">
> > <manager-properties>
> > <property name="relaxCacheVersionSemantics" value="true" />
> > </manager-properties>
> > </session-manager>
> > </session-config> 
> > 
> 
> 
> As I think you probably understand, the session referred to there is not
> anything to do with the SP. It doesn't know about Java and it can't
> connect up with that session.
> 
> Perhaps there's a Glassfish bug related to header size and populating all
> those headers causes it to react badly. If using the environment works, as
> it should with any AJP-capable container, then it seems like that's your
> workaround.
> 
> -- Scott
> 
> 
> --
> To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net (mailto:users-unsubscribe at shibboleth.net)
> 
> 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shibboleth.net/pipermail/users/attachments/20130903/b239c86f/attachment.html