Firing plug shows wite content area

Question

Hello everybody!
Currently we encounter a strange problem in one of our WebDynpro applications. When a certain user clicks on a button that fires a plug to navigate to another view, the contet area gets white and nothing else happens. This happens to this user only, it happens whenever the navigation t a certain view is triggered. For every other user, this navigation works perfectly, even with the same data loaded. Another user experiences similar behaviour when navigation to another view; again, for every other user the navigation works fine.
When entering any view, we check if the user is member of a certain UME group. But this check is performed identically (for the same UME group) in every view and that works fine in every case except for those two user-view-combinations.
Did anyone come across something like that or have any idea on what might be going wrong ? I'd happily supply any information necessary...
Thanks
Michael

Answer 1

Hi!
I had a look at the SCA dependencies of WDRUNTIME12P_11-20007087.SCA, and found that it requires AJAX RUNTIME 7.31 with minimum patch level 6 (ours is 3), FRAMEWORK EXTENSIONS 7.31 with patch level 6 (ours is 3) and PORTAL BASIS 7.31 wih minimum patch level 12, while ours is 7. I'll try to make our basis team install the latest patches on everything, even if that does not help with my problem.
Yes, there are only two users (of about 30 to 40 that use this application) having that problem, each of them when navigating to another view, but the white content area displays everytime these users try to navigate to their "special" view. Clearing the browser cache did not help, and the problem remains even when using Chrome instead of IE9. Both users are working on terminal services (Citrix), currently we are checking if something changes when they use ordinary Windows 7 workstations.
Funny enough, the application initially worked fine for both users, the behaviour changed some weeks ago for the first user and a few days ago for the other one. As far as I know, there was no change in portal configuration, no patches were applied and we did not deploy another version of our application at these times. And this happens on our production machine, only, of course.
Since we are checking UME memberships when navigating to any view, I was wondering if there is a circular reference in the UME group or role definition, but since we are checking for membership in the same group in every view using the same class, I can hardly imagine why this should fail in one view only (and a different one for both users).
Anyway, if anyone might come up with another idea, I'd be grateful...
Thanks Michael