OK. I already created one for the easy part of the NPE.
Susan
Eric Dalquist wrote:
Susan,
Can you file a jira issue with all of the information you have below
(and maybe some of those stack traces?)
Thanks,
-Eric
Bramhall, Susan wrote:
I am seeing some interesting behavior in
3.1. I believe it occurs when I hit the portal with a url from an
expired session. For example
http://portal.../uPortal/render.userLayoutRootNode.uP?uP_root=root&uP_sparam=activeTab&activeTab=2.
I have added the CAS gateway=true parameter to my web.xml and haven't
yet determined whether the problem still exists when that is not set.
Instead of displaying the guest page I get several problems. The first
one is a stack trace sent to the browser:
Caused by: java.lang.NullPointerException
at
org.jasig.portal.rendering.StaticRenderingPipeline.renderState(StaticRenderingPipeline.java:412)
That is actually trying to log an error with a null cause. So I change
that line to not use e.getCause() and just print the error and get:
ERROR [http-443-5] portal.ChannelManager.[] Mar/13 06:47:44 -
Encountered a portal exception while trying to start channel rendering!
:
org.jasig.portal.PortalException: Could not find a transient node for
id: u7l1n10
That's not too too surprising since the old session was dead I suppose
but it doesn't clear up. Once this problem occurs it does not go
away. Even if I hit the portal from another browser on another machine
the error behavioscontinues. I have lots more stack traces but will
spare you all for now. ;-) The screen shots showing tha uportal
thinks I am editing a fragment tell more. When I use the normal url to
the portal with no parameters I see the image attached as
up3expiredguestnoparms.jpg. The second is when I hit tab2 and is
attached as up3expiredguesttab2.jpg.
It looks to me like I am running into issues related to the new
impersonation support. It gets even more interesting if I actually get
myself a CAS session - then I get two messages to exit the fragment.
Image attached as up3expiredguestCASsession.jpg.
I suspect the Could not find a transient node problem has been around a
long time since our logs are full of an older varient which I bet is
due to CAS gateway and guest timeouts. It used to report the node as
ctf1 and the new message reports more meaningfull data. Our old error
is completely transient and has no effect on the user experience as far
as I can tell.
I am looking forward to debugging this (really) but though I'd post to
get thoughts from you all first.
Susan
--
Susan Bramhall (susan.bramh...@yale.edu)
Senior Developer, Technology & Planning Operations (T&P)
Yale University Information Technology Services (ITS)
25 Science Park, 150 Munson St, New Haven, CT 06520
Phone: 203 432 6697
--
You are currently subscribed to uportal-dev@lists.ja-sig.org as: arch...@mail-archive.com
To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/uportal-dev
|
