Apache 2.4 producing 500 error when serving DirectoryIndex page from Shib SP protected areas

Cantor, Scott cantor.2 at osu.edu
Wed Oct 3 11:31:18 EDT 2012


On 10/3/12 10:38 AM, "Paul Beckett (ITCS)" <P.Beckett at uea.ac.uk> wrote:
>
>Removing the hook to /AttrChecker, doesn't affect the behaviour (in terms
>of the end result in the browser). I'm pretty sure it's passing though
>the AttrChecker successfully, as it initially failed displaying an error
>template (due to our not releasing a required field in our IDP). When we
>released that field it now progresses to the " Please wait while we
>transfer you to 
>https://ueacisr6test.uea.ac.uk/Shibboleth.sso/Metadata..." (my SP), the
>same as if I remove the sessionHook entirely. My expectation was that if
>the client requested https://ueacisr6test.uea.ac.uk/secure/ then they
>would be authenticated and sent back to the original location
>(https://ueacisr6test.uea.ac.uk/secure/) to see the content.

Yes, correct. All that fits.

>After the message is displayed, if I re-request the page:
>https://ueacisr6test.uea.ac.uk/secure/ the browser whirrs away, but
>nothing gets returned.

Ok. Is there anything at that location or in that directory? Just want to
make sure I can reproduce the scenario exactly so I can debug it.

The basic problem is that Apache doesn't process subrequests properly
(defined as "using the same sequence of module hooks and constraints as
the original request"). I have to figure out what each case looks like and
account for them.

>Sorry the access log details I included were just the bit after the
>client returned from the IDP, when no sessionHook was configured. Access
>log details of the different sessionHook configurations is below.

That's all expected. It's obviously hung in the /secure request in both
cases. So the hook's not involved, and is working fine.

-- Scott




More information about the users mailing list