Windows Server 2008 Installation—It’s a Snap!

Downloads
98081.zip

The Windows Server 2008 installation process is a very different beast than what you’ve experienced in the past when you’ve rolled out a new version of Windows Server. On the surface, the installation process might appear to be similar to what you’ve done before—but with fewer questions to answer—when in fact, something very different is happening under the covers. The speed of the installation will tip you off: It’s quick!

Like Windows Vista, Server 2008 is an image-based installation from a Windows Imaging Format (WIM) file on the Server 2008 DVD. And just as the Vista DVD contains all the versions of that OS (i.e., Home Basic, Home Premium, Ultimate, and Business), the Server 2008 DVD contains the main versions of Server 2008: Windows Server 2008 Standard, Windows Server 2008 Enterprise, and Windows Server 2008 Datacenter editions. All the versions fit on one DVD because WIM is a single-instance storage format. That is, because the various versions contain the same basic set of files, all the versions can be stored in one image that takes up only slightly more space than the image of a single version. Each Server 2008 DVD supports only one architecture, so you’ll have a different DVD for x86 (32-bit) than for x64 (64-bit). Let’s walk through a typical installation scenario.

What to Expect
When you install Server 2008, you first need to decide whether to perform a clean installation or an upgrade. Usually, a clean installation is the best option, and that’s what this example will show. (If you choose to upgrade, see the sidebar “What You Need to Know About In-Place Upgrades.”)

You can deploy Server 2008 using Windows Deployment Services (WDS), which sends the installation environment over the network and lets you easily automate configuration with an unattended answer file. But to really see the ins and outs of the installation process, let’s install Windows Server 2008 the old-fashioned way—manually.

Insert the Server 2008 DVD into your system’s optical drive, and choose to boot from media (i.e., the DVD), which will load the Windows Preinstallation Environment (WinPE) from the boot.wim file on the DVD. Because Server 2008 is an image-based installation, the system needs an environment on which to lay the image, in addition to other functionality (e.g., the capability to partition the hard disk). WinPE provides that environment.

Once WinPE loads onto the system, the installation process immediately makes sure that the system has at least 512MB of memory; if the system has less memory, the installation won’t proceed. If the system has enough memory, the installation process prompts you to select the language, time and currency formats, and keyboard or input method you want to use in the installation. The default is U.S. English, but you can modify the settings to fit your environment.

Next, you see a window that gives you the option to “Install now” or “Repair your computer.” Selecting “Install now” will launch the installation routine, setup.exe. At any time while setup.exe is running, you can press Shift+F10 to open a command window in case you need to perform any other functions, such as running a script to add a utility partition or troubleshoot a problem installation; as long as the command window is open, the installation routine won’t reboot the server. The repair option provides access to the Windows Recovery Environment and some automated repair options. These repair capabilities are very useful, so it’s a good idea to keep the Server 2008 DVD handy for future use. (Note that you can also create a repair disk at any time from within Server 2008 after you install the Windows Backup Server feature; creating a repair disk is an option of the backup feature.)

So click “Install now,” and the installation process displays a window that asks you to enter your 25-character product key, which is linked to a specific version of Server 2008. You can enter the product key, or you can leave the field blank and just click Next, which will trigger the confirmation dialog box that you see in Figure 1.

Why might you choose not to enter your product key at this time? Maybe you want to test the OS for 30 days, or maybe you just prefer to copy and paste the product key from a file after you install the OS. When you activate Server 2008, which you must do within 30 days after installing the OS, just be sure that the version installed on your system matches the version you’re licensed for. If not, you’ll face two options: You can purchase a product key for the version installed on your machine (which could be costly if, for example, your installed version is the Enterprise edition and your product key is for the Standard edition), or you can reinstall the version that matches your product key (which might cause you to lose any data, information, and programs placed on your system since you last installed the OS).

For this example, click No in the confirmation dialog box. A new window opens and asks which edition of Server 2008 you want to install. Because you didn’t enter a product key previously, the window displays all the versions of Server 2008 that are in the image file, along with a confirmation check box that states, “I have selected the edition of Windows that I purchased,” as Figure 2 shows.

If you had previously entered a product key, you’d see just two versions of the OS to choose from: the Full Installation and the Server Core Installation versions for the edition of the OS the product key identified. Server Core is a “lite,” minimal-footprint server installation option that provides a lowmaintenance, limited-functionality server environment. Server Core offers only basic components of Server 2008—not even the Windows Explorer shell, but just a commandprompt user interface—and is capable of supporting core server roles such as file server, DHCP server, print server, and DNS server. Server Core is not a platform for application development or application serving, for example, because it doesn’t include the .NET Framework. So, why would you choose the Server Core option? With only minimal Windows functionality, many product updates will not need to be applied to Server Core systems; consequently, Server Core systems require less maintenance. Server Core systems also use fewer resources (e.g., disk space), and with no GUI, they’re less open to security risks.

Once you select the Server 2008 version you want to install and click Next, the licensing agreement is displayed. As always, read it thoroughly to ensure that you agree to all the conditions, select the “I accept the license terms” check box, and click Next.

Continued on page 2

The next window prompts for the type of installation you want to perform: Upgrade or Custom (advanced). Because you’re doing a clean installation from media, the Upgrade option is disabled (i.e., grayed out), and you must select Custom (advanced). Note that if you run the installation process from within Windows Server 2003, both options will be enabled.

A new window opens that asks one final question: Where do you want to install Windows? A dialog box displays partitions and unallocated space. You can add or remove partitions, reformat a previously used hard drive before installing Server 2008, and load additional drivers as needed. Select the partition for the installation, then click Next to install the OS. If the partition isn’t formatted, the installation process quickly formats the partition as NTFS and proceeds.

That’s it. You’re done—no more questions. You can go have a drink. But don’t go too far! Because the installation is image-based, it doesn’t take long to complete. A window displays the progress of the installation, and the server reboots twice during the process.

Configuration
So, what about all the things you never configured during installation: server name, time zone, administrator password, IP configuration? A server has a lot of default settings, DHCP-assigned IP address, automatically assigned server name, and so forth that you need to configure after the installation process has completed. This is starting to sound worse than what you had to do previously to install Windows Server! In the past, you installed and configured the OS in one process. Now, do you have to root through different Control Panel applets to configure the server? Fortunately, no.

As Figure 3 shows, the first thing you see after the installation is a window that tells you the user’s password must be changed before logging on for the first time. In the Server 2008 installation process, the Administrator account is created with a blank password, so the first action is to set a new Administrator password. Once you’ve set the new Administrator password, you are logged on as the Administrator.

If you ever had to install Windows Server 2003 with Service Pack 1 (SP1) slipstreamed after installation, you know that the Post Setup Security Updates (PSSU) Wizard forced you to patch your server and set an update schedule. With Server 2008, you get a beefed-up PSSU-type process in the form of the Initial Configuration Tasks (ICT) interface.

As you can see in Figure 4, the ICT guides you through all the main configuration items for a server with a new installation of Server 2008. The current values are displayed, and clicking an item opens the appropriate Control Panel applet for the value you want to set. For example, when you click the icon to set the computer name and domain, the Control Panel System applet opens automatically.

In Server 2008, Windows Firewall is enabled and Remote Desktop is disabled by default so that the server is secure from the start. Furthermore, Windows Firewall is fully integrated with the OS. Server 2008 offers several server roles (e.g., DHCP server, DNS server, domain controller) and features (e.g., backup, clustering) that help the server perform the role you select. (You add roles and features in section 3 of the ICT interface via the “Add roles” and “Add features” links.) When you enable a role and its supporting features, the various ports required by the role and its features are opened automatically in Windows Firewall; no additional configuration is required. (To maintain the Windows Firewall settings over time, you’ll want to use the Security Configuration Wizard—SCW— to create templates that let you continuously monitor Windows Firewall.)

Once you’ve configured the server, select the “Do not show this window at logon” check box and click Close. If you want to perform further configuration or role and feature maintenance, use Server 2008’s new role-based management tool, Server Manager.

Beyond Manual Installation
So that’s a walkthrough of the basic Server 2008 installation experience. As you’ve seen, you don’t really have much to do, but unless you need to install the OS on just a few servers, you’ll want to automate the installation process.

To create an unattended answer file for use in an automated process, first download Microsoft’s free Windows Automated Installation Kit (www.microsoft.com/downloads/details.aspx?FamilyID=c7d4bc6d-15f3-4284-
9123-679830d629f2&DisplayLang=en). The WAIK contains the Windows System Image Manager application, which you’ll use to create your answer file. You can use the answer file with services such as Windows Deployment Services to automate your installations, or you can name the answer file autounattend.xml, place it on a floppy disk or USB drive, and insert it as the Server 2008 installation process begins. The process will read and use the answer file to automate the installation.

The WAIK documentation details the minimum requirements you need to specify for an automated installation. Web Listing 1 (Download a .zip file at the top of the page this page) provides an example of an autounattend. xml file that will partition the disk and install the full version of Server 2008 Enterprise. To use this file, you need to set the product key and also the local Administrator password value via the Windows System Image Manager as the local Administrator password is encrypted. (Note that there are other options—for example, a key management system—that don’t require you to hand out the product key in an autounattend file.)