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]> 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]> 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**** >> >> 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 **** >> >> ** ** >> >> *From:* Karl Wright [mailto:[email protected]] >> *Sent:* Thursday, May 16, 2013 12:05 PM >> *To:* [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]> 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]> 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 **** >> >> **** >> >> 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) **** >> >> **** >> >> ** ** >> >> ** ** >> > >
