At work I've been doing validation paper work (see https://www.ntldr.net/Lists/Blog/DispForm.aspx?ID=11), but that changed this Monday.  We've had some "productivity committees" working on making the company more productive since the spring.  Apparently, the latest finding of one of these groups is that we need better access/search to old reports and data summaries that are stored on one of the servers.  Ut-oh...that means these silly (waste of time, imho, since they don't seem to be addressing the real issues) things now affect IT.

 

So I've been reprioritized to be SharePoint-Intranet-Search development guy.  I started working on the intranet site during the holiday season last year, as a low-priority, "do this when you've got nothing better to do, like reimage perfectly good PC's", type of task.  Prior to my taking over, we had a basic intranet site that didn't work, and wasn't being worked on.  So my first task was to bring it forward 4 years and sit it on modern technologies.  Which is part of the reason I chose SharePoint.  I also chose it because I hadn't worked with it before, but had seen some interesting demos, heard some stuff about it, and figured I'd treat it as a learning experience.  As it turns out, SharePoint also had just about everything we needed in a Intranet site builtin, but the IT Director didn't like the UI, so I customized that a bunch, and got it looking decent.

 

So everything was going along great with the site and the test server it was on (our production webserver was still on NT4 with IIS4 limping along), until an instrument PC died.  The instrument actually makes money for the company, so I made the call and the test pc got appropriated.  This wouldn't have been a problem, except that the backup I shot of the test system, including site config, etc. turned out to be corrupted when I went to restore it after the instrument pc problem was permanently resolved.  I know, I know, VERIFY backups...

 

In any case, shortly thereafter it came time to upgrade the servers, and then do the validation, so work on the site got pushed back a bunch.  About a month ago I started work back up on the site, basically getting it back to where I was before the whole loss of the pc.  This included getting  the few custom apps I'd written (software inventory search, computer location on map, and basic document search) back up and running.  The one issue that I ran into here was that Kerberos seemed to not be working...impersonation was always failing when connecting to the db server.

 

It turns out Kerberos is completly broken in our AD domain.  I HAVE NO IDEA HOW THIS HAPPENED.  The network admin is going through the troubleshooting guides now and should have some answer for me soon.  In the meantime, it's back to Basic auth and debugging of the search stuff.

 

The current search mechanism uses the Indexng Service builtin to Windows Server 2003, and seems to do a fair job providing search results.  Large result sets do cause perf to take a bit of a hit, but I'm hoping that those wont be the normal search types.  Profiling and focus group testing should help clarify the extent of this problem.  The biggest problem isn't going to be coding related though, it's going to be dealing with the management personel who are advocating the search, since I get the feeling that this is more a reaction to the whole search phenomena going around the web right now.  Time to go get specs locked in...(but that needs to wait until finals week is over).