Core to any server is the ability to service networks requests. In some cases, raw I/O is the goal. In other cases, simultaneous streams might be the goal. Windows Server 2008 Core is the engine for many of those cases. Do you need a graphical user interface to run a server? Not really so lets see what this new Core implementations is all about.
The installation of Windows Server 2008 Core is nearly identical to a regular server install. However, the end result is VERY different. As with the previous screencast and install, we supply the product key, answer a few questions and we're off. The Core installation is much quicker because there's less "stuff" to install. This benefits you in a variety of ways. Core consumes less disk space. That's really obvious. Since Core is a much smaller set of applications, processes and services, the potential attack surface or vulnerability landscape is much smaller.
[errata note] I made a mistake at the tail end of the setup portion of this screencast. I said we were going to promote the Core installation to a DC, then to a RODC. That is incorrect. We will convert directly to a Read Only DC (RODC).
The best place to get information on how to setup and configure Windows Server 2008 Core is of course the Step-by-Step guide. You'll learn a lot from this guide but of course doing is better than reading.
As indicated in the screencast below, one of the first things you'll want to do is set the administrator password. The guide shows the command line method. Regardless of the method used, do it, do it fast.
When you install the Core server, a machine name is generated and it isn't pretty like the one suggested by Windows Vista. It starts with LH- followed by a nice string of characters. You can of course create the machine name at setup if you are driving the setup process with an unattended installation script. Unattended installation is pretty easy but for the purposes of this screencast, we'll defer that magic. Make sure to use the following command to rename your server:
netdom renamecomputer %computername% /newname:<NewComputerName>
You can certainly use the command in the guide, but that means you need the generated machine name and I'm lazy. The command above will grab the machine name from the %computername% variable. Machine name changes require a reboot so you may want to hold off on that until you configure the network interfaces.
To activate or not to activate, that is the question. I would imagine you'll be testing and learning from your installation for longer than 30 days. If that's the case, you must activate. Activation is easy enough. The following command assumes you have network connectivity to the Microsoft activation servers.
Slmgr is a .vbs script present in Windows Vista and Windows Server 2008. It has a number of useful command line arguments. For instance, the -xpr command will tell you if your license period is about to expire.
Another useful argument is -ipk. It comes in handy if you burn through all of the activations for a particular key and need to change it. -rearm extends the grace period, but you can only rearm Windows a finite number of times.
The one role or service installed by setup is the file server service. You can bring up a computer manager and connect to the Core server and immediately create a share and start using it for file sharing purposes. Some configuration of the firewall will likely be necessary for some of your designs.
Installing roles is pretty easy. Make sure to remember the role names are case sensitive to the installation tools. I know, that's odd. In the screencast, we install the FRS-Infrastructure role. There are a number of other possible roles.
Role installation allows your Core server to potentially specialize. By specialize, I mean you could strategically position certain Core servers on your network to handle specific types of demand. This may seem contrary to the consolidation trend over the years, but also keep in mind that a Core server can run multiple roles, too. The bottom line is that it's flexible.
Just like it's big GUI brother, the Core server can also install a number of features. Those features include Failover Clustering, Network Load Balancing, Subsystem for UNIX-based applications, Backup and others. Some of the features are only available in the Enterprise Core option.
I'll be posting a screencast soon on how to convert a Windows Server 2008 Enterprise Core into a Read Only Domain Controller (RODC) and filter the authentication password list. It's a very interesting demo so keep your eyes peeled for it.
This screencast is longer than the previous screencasts. I combined the setup screencast with the configuration details screencast. It's still pretty short at 18 minutes. Let me know if you prefer smaller or bigger chunks of demo video. The more complex the demonstration, the more I'll have to capture so the topics will naturally start to get longer. Here's the direct link to the Windows Server 2008 Core Setup and Configuration screencast:
http://forums.microsoft.com/TechNet/ShowForum.aspx?ForumID=582&SiteID=17 - Server Core TechNet forum