Yeah, I just noticed I have java issues, which is where my problems are coming 
in. Thanks.

TC Tobin-Campbell | Technical Services | Willow | Epic  | (608) 271-9000

From: Karl Wright [mailto:[email protected]]
Sent: Friday, May 31, 2013 10:02 AM
To: [email protected]
Subject: Re: ManifoldCF and Kerberos/Basic Authentication

You can look at the how-to-build-and-deploy page, but basically this is what 
you need to do:

ant make-core-deps
ant build
... then you will find everything the same as you are familiar with under the 
dist directory.

Karl

On Fri, May 31, 2013 at 10:59 AM, TC Tobin-Campbell 
<[email protected]<mailto:[email protected]>> wrote:
Hey Karl,
This is probably a dumb question, so consider yourself warned.

I got the trunk downloaded and built, but now can't figure out how I'm supposed 
to run ManifoldCF from it. In the prebuilt downloads, there are example 
directories where I can go in and just click the start.jar file and it all 
kicks off fine. Not in the trunk. Any suggestions?

TC Tobin-Campbell | Technical Services | Willow | Epic  | (608) 
271-9000<tel:%28608%29%20271-9000>

From: Karl Wright [mailto:[email protected]<mailto:[email protected]>]
Sent: Friday, May 24, 2013 12:50 PM

To: [email protected]<mailto:[email protected]>
Subject: Re: ManifoldCF and Kerberos/Basic Authentication

I had a second so I finished this.  Trunk now has support for basic auth.  You 
enter the credentials on the server tab underneath the API credentials.  Please 
give it a try and let me know if it works for you.

Karl

On Fri, May 24, 2013 at 11:28 AM, Karl Wright 
<[email protected]<mailto:[email protected]>> wrote:
CONNECTORS-692.  I will probably look at this over the weekend.
Karl

On Fri, May 24, 2013 at 11:26 AM, Karl Wright 
<[email protected]<mailto:[email protected]>> wrote:
Hi TC,
Unless I'm very much mistaken, there are no Apache kerberos session cookies 
being used on your site, so it should be a straightforward matter to include 
basic auth credentials to your Apache mod-auth-kerb module for all pages during 
crawling.
I'll create a ticket for this.

Karl

On Fri, May 24, 2013 at 11:14 AM, TC Tobin-Campbell 
<[email protected]<mailto:[email protected]>> wrote:
Hi Karl,
Here's what I know so far.

Our module is configured to use two auth methods: Negotiate and Basic.  In most 
cases, we use Negotiate, but I'm guessing you'd prefer Basic.

Here's an example header.

GET / HTTP/1.1
Host: wiki.epic.com<http://wiki.epic.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:20.0) Gecko/20100101 
Firefox/20.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate
Cookie: wooTracker=QOMVLXDIC6OGOUXMGST1O54HYW573NNC; 
.EPICASPXAUTHQA=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;
 wiki_pensieve_session=j1pcf1746js1442m7p92hag9g1; wiki_pensieveUserID=5; 
wiki_pensieveUserName=Lziobro; 
wiki_pensieveToken=********************be3a3a990a8a
Connection: keep-alive
Authorization: Basic bHppb**************xMjM0   <-I've censored this line so 
you cannot get my password

If I'm understanding you correctly, there's no way to accomplish this 
currently? Or, is there some workaround we could implement?

TC Tobin-Campbell | Technical Services | Willow | Epic  | (608) 
271-9000<tel:%28608%29%20271-9000>

From: Karl Wright [mailto:[email protected]<mailto:[email protected]>]
Sent: Thursday, May 16, 2013 12:05 PM
To: [email protected]<mailto:[email protected]>
Subject: Re: ManifoldCF and Kerberos/Basic Authentication

Hi TC,

Apparently mod-auth-kerb can be configured in a number of different ways.  But 
if yours will work with basic auth, we can just transmit the credentials each 
time.  It will be relatively slow because mod-auth-kerb will then need to talk 
to the kdc on each page fetch, but it should work.  Better yet would be if 
Apache set a browser cookie containing your tickets, which it knew how to 
interpret if returned - but I don't see any Google evidence that mod-auth-kerb 
is capable of that.  But either of these two approaches we could readily 
implement.
FWIW, the standard way to work with kerberos is for you to actually have 
tickets already kinit'd and installed on your machine.  Your browser then picks 
up those tickets and transmits them to the Wiki server (I presume in a header 
that mod-auth-kerb knows about), and the kdc does not need to be involved.  But 
initializing that kind of ticket store, and managing the associated kinit 
requests when necessary, are beyond the scope of any connector we've so far 
done, so if we had to go that way, that would effectively make this proposal a 
Research Project.
What would be great to know in advance is how exactly your browser interacts 
with your Apache server.  Are you familiar with the process of getting a packet 
dump?  You'd use a tool like tcpdump (Unix) or wireshark (windows) in order to 
capture the packet traffic between a browser session and your Apache server, to 
see exactly what is happening.  Start by shutting down all your browser 
windows, so there is no in-memory state, and then start the capture and browse 
to a part of the wiki that is secured by mod-auth-kerb.  We'd want to see if 
cookies get set, or if any special headers get transmitted by your browser 
(other than the standard Basic Auth "Authentication" headers).  If the exchange 
is protected by SSL, then you'll have to use FireFox and use a plugin called 
LiveHeaders to see what is going on instead.
Please let me know what you find.
Karl


On Thu, May 16, 2013 at 12:37 PM, Karl Wright 
<[email protected]<mailto:[email protected]>> wrote:
Hi TC,
Thanks, this is a big help in understanding your setup.
I don't know enough about exactly *how* mod-auth-kerb uses Basic Auth to 
communicate with the browser, and whether it expects the browser to cache the 
resulting tickets (in cookies?)  I will have to do some research and get back 
to you on that.
Basically, security for a Wiki is usually handled by the Wiki, but since you've 
put added auth in front of it by going through mod-auth-kerb, it's something 
that the Wiki connector would have to understand (and emulate your browser) in 
order to implement.  So it does not likely support this right now.  It may be 
relatively easy to do or it may be a challenge - we'll see.  I would also be 
somewhat concerned that it may not possible to actually reach the API urls 
through Apache; that would make everything moot if it were true.  Could you 
confirm that you can visit API urls through your Apache setup?
Karl

On Thu, May 16, 2013 at 12:21 PM, TC Tobin-Campbell 
<[email protected]<mailto:[email protected]>> wrote:
Hi there,
I'm trying to connect ManifoldCF to an internal wiki at my company. The 
ManifoldCF wiki connector supplies a username and password field for the wiki 
api, however, at my company, a username and password is required to connect to 
the apache server running the wiki site, and after that authentication takes 
place, those credentials are passed on to the wiki api.

So, essentially, I need a way to have ManifoldCF pass my windows credentials on 
when trying to make its connection. Using the api login fields does not work.

We use Kerberos the Kerberos Module for 
Apache<http://modauthkerb.sourceforge.net/index.html> (AuthType Kerberos).  My 
understanding based on that linked documentation is that this module does use 
Basic Auth to communicate with the browser.

Is there anything we can to make ManifoldCF authenticate in this scenario?

Thanks,


TC Tobin-Campbell | Technical Services | Willow | Epic  | (608) 
271-9000<tel:%28608%29%20271-9000>

Sherlock<https://sherlock.epic.com/> (Issue tracking)
Analyst 
Toolkits<https://sites.epic.com/epiclib/epicdoc/Pages/analyst/default.aspx> 
(Common setup and support tasks)
Report 
Repository<https://documentation.epic.com/DataHandbook/Reports/ReportSearch> 
(Epic reports documentation)
Nova<https://nova.epic.com/Login/GetOrg.aspx?returnUrl=%2fdefault.aspx> 
(Release note management)
Galaxy<https://documentation.epic.com/OnlineDoc/Documents.aspx> (Epic 
documentation)







Reply via email to