The Pedantic Details of Websites Between Servers
by Mike Levin SEO & Datamaster, 05/02/2009
OK, now for moving the websites off of Brain. There are only two that appear to be in use and important. One is web.connors.com, which is where all the tracking GIF systems that preceeded HitTail resolved to. We used the client subdomain DNS entry trick to make it part of their network, and avoid original origin domain security warnings in the browser. Since then, we moved to P3P policies which has the same effect without all the technical
Anyway, in moving Websites in vintage Windows Server 2003 / IIS6 software, there are really two steps. The first is going into IIS through the Microsoft Management Console, right-clicking a site, slecting “All Tasks” and then selecting “Save Configuration to a File”. That gets all the difficult-to-reproduce IIS Website configuration stuff into a loadable file. The second part of this is actually copying the files using the normal File Exporer. I have to wait for the 7GB file copy to finish so as to not cause file-write contention. Hum dee dum.
A potential third step of moving a Website on IIS is moving any registered server components that the websites may be dependent upon. Back in the olden days, I used a compiled Visual Basic 6 object to acheive the serving of the tracking GIF. This was the documented way to do it back in say 1999 when I first started researching “the Web bug” that makes all this tracking stuff possible. In later days, I switched over to the built-in ADODB object that accomplisehes exactly the same thing without dependencies on specialty components. I forget which way this tracking system is set up. I need to investigate that now, because if it’s the stupid DLL way, I need to move that file too, and remember the component registration process for later tonight.