Preplanning and Preparing a Server Installation

Before you begin the actual installation of Windows Server 2003, you must make several decisions. How well you plan these steps will determine how successful your installation is.

Verifying Minimum Hardware Requirements

The first step of the installation is verifying that your hardware meets the system requirements. Keep in mind that, although there is a minimum requirement for the CPU and RAM, there is also a recommended CPU and RAM configuration. For the sake of performance, you should usually stay away from the minimum requirements and stick to the recommended settings (or better). Table 3.1 lists system recommendations for Windows Server 2003 (32-bit), and Table 3.2 lists system recommendations for Windows Server 2003 (x64-bit edition).

Table 3.1. System Requirements (32-bit)
Requirement	Standard Server	Enterprise Server
Minimum CPU speed	133MHz	133MHz for x86-based computers
		733MHz for Itanium-based computers
Recommended CPU speed	550MHz	733MHz
Minimum RAM	128MB	128MB
Maximum RAM	4GB	32GB
Multiprocessor support	Up to 4	Up to 8
Disk space	1.5GB	1.5GB

Table 3.2. System Requirements (x64-bit)
Requirement	Standard Server	Enterprise Server
Supported Processors	AMD Opteron, AMD Athlon 64, Intel Xeon with Intel EM64T, and Intel Pentium with Intel EM64T
Minimum RAM	512MB	512MB
Maximum RAM	32GB	1TB
Multiprocessor support	Up to 4	Up to 8
Disk space	4GB	4GB

Choosing a New Installation or an Upgrade

If you have an existing Windows environment, you may need to perform a new installation or upgrade an existing server. There are benefits to each of these options.

Should You Perform a New Installation?

The primary benefit of a new installation is that, by installing the operating system from scratch, you are starting with a known good server. You can avoid migrating problems that may have existed on your previous serverwhether due to corrupt software, incorrect configuration settings, or improperly installed applications. Keep in mind, however, that you will also lose all configuration settings from your previous installation. Make sure you document your server configuration information and back up any data that you want to keep.

When performing a new installation, you can install on a new hard drive (or partition) or in a different directory on the same disk as a previous installation. Most new installations are installed on a new or freshly formatted hard drive. Doing so removes any old software and gives you the cleanest installation.

Should You Upgrade an Existing Server?

Upgrading, on the other hand, replaces your current Windows files but keeps existing users, settings, groups, rights, and permissions. In this scenario, you don't have to reinstall applications or restore data. Before choosing this option, keep in mind that you should test your applications for compatibility before migration. Just because they worked on previous versions of Windows does not mean they will work on Windows Server 2003.

As always, before performing any type of server maintenance, you should perform a complete backup of any applications and data that you want to preserve.

To upgrade to Windows Server 2003, you must be running a server-level operating system. You cannot upgrade Workstation or Home editions to Windows Server 2003. To upgrade your existing server, you must be running Windows 2000 or Windows NT 4.0 Server (Service Pack 5 or higher). Table 3.3 lists the available upgrade paths to Windows Server 2003.

Table 3.3. Upgrade Compatibility for Windows Server 2003
Previous Operating System	Upgrade to Windows Server 2003 Possible?
Windows NT versions 3.51 and earlier	No, you must first upgrade to NT 4.0 Service Pack 5 or higher.
Windows NT 4.0 Server	Yes, you must have Service Pack 5 or higher.
Windows 2000 Server	Yes.
Windows 2000 Advanced Server	Yes.
Windows 2000 Professional	No, only server-level operating systems can be upgraded.
Windows XP Professional	No, only server-level operating systems can be upgraded.
Novell NetWare	No, but migration tools are available to migrate Novell Directory Services (NDS) information to a Windows domain.

Note

When considering installing or updating to Windows 2003 R2, the R2 update merely adds components on top of an existing Windows 2003 SP1 or higher server configuration. Windows 2003 R2 is not a new or separate operating system. Therefore, when referencing the upgrade of Windows, this chapter refers to the upgrade from Windows NT4 or Windows 2000 to Windows 2003, and not from Windows 2003 to Windows 2003 R2.

Determining the Type of Server to Install

You have the choice of making your server a domain controller (DC), a member server, or a standalone server. After you determine the tasks the server will perform, you can determine the role you will assign to it.

Domain controllers and member servers play a role in a new or existing domain. Standalone servers are not joined to a particular domain.

As in Windows 2000, you are able to promote or demote server functions as you like. Standalone servers can be joined to the domain to become member servers. Using the DCPromo utility, you can promote member servers to domain controllers. And, by uninstalling the Active Directory service from a domain controller, you can return it to member server status.

Gathering the Information Necessary to Proceed

During the installation of Windows Server 2003, you will have to tell the Setup Wizard how you want your server configured. The wizard will take the information you provide and will configure the server settings to meet your specifications.

Taking the time to gather the information described in the following sections before starting your installation will likely make your installation go faster and easier.

Selecting the Computer Name

Each computer on a network must have a name that is unique within that network. Many companies have a standard naming convention for their servers and workstations. If not, you can use the following information as a guideline for creating your own. Although the computer name can contain up to 63 characters, workstations and servers that are preWindows 2000 recognize only the first 15 characters. It is widely considered a best practice to use only Internet-standard characters in your computer name. This includes the letters AZ (upper- and lowercase), the numbers 09, and the hyphen (-).

Although it's true that implementing the Microsoft domain name system (DNS) service in your environment could allow you to use some non-Internet standard characters (such as Unicode characters and the underscore), you should keep in mind that this is likely to cause problems with any non-Microsoft DNS servers on your network. You should think carefully and test thoroughly before straying from the standard Internet characters noted in the preceding paragraph.

Name of the Workgroup or Domain

During the server installation, the Setup Wizard will ask for the name of the workgroup or domain that the server will be joining. You can either enter the name of an existing organizational structure or enter a new name, creating a new workgroup or domain.

Users new to Microsoft networking may ask, "What is the difference between a workgroup and a domain?" Simply put, a domain is a collection of computers and supporting hardware that share the same security database. Grouping the equipment in this manner allows you to set up centralized security and administration. Conversely, a workgroup has no centralized security or administration. Each server or workstation is configured independently and locally for all security and administration settings.

Network Protocol and IP Address of the Server

When installing Windows Server 2003, you must install and configure a network protocol that will allow it to communicate with other machines on the network. Currently, the most commonly used protocol is called TCP/IP, which stands for Transmission Control Protocol/Internet Protocol. This protocol allows computers throughout the Internet to communicate. After you install the TCP/IP protocol, you need to configure an IP address for the server. You can choose one of the following three methods to assign an IP address:

Automatic Private IP Addressing (APIPA) APIPA can be used if you have a small network that does not have a Dynamic Host Configuration Protocol (DHCP) server, which is used for dynamic IP addresses. A unique IP address is assigned to the network adapter using the LINKLOCAL IP address space. The address always starts with 169.254 and is in the format 169.254.x.x. Note that if an APIPA is in use, and a DHCP server is brought up on the network, the computer will detect this and will use the address that is assigned by the DHCP service instead.
Dynamic IP address A dynamic IP address is assigned by a DHCP server. This allows a server to assign IP addresses and configuration information to clients. Some examples of the information that is distributed include IP address, subnet mask, default gateway, domain name system (DNS) server address, and Windows Internet Naming Service (WINS) server address. As the dynamic portion of the name suggests, this address is assigned to the computer for a configurable length of time, known as a lease. When the lease expires, the workstation must again request an IP address from the DHCP server. It may or may not get the same address that it had previously. Although servers and workstations can both be configured to use this method of addressing, it is generally used for workstations rather than servers.
Static IP address Using a static IP address is the most common decision for a server configuration. By static, we mean that the address will not change unless you change the configuration of the server. This point is important because clients and resources that need to access the server must know the address to be able to connect to it. If the IP address changed regularly, connecting to it would be difficult.

Backing Up Files

Whether you are performing a new installation on a previously used server or upgrading an existing server, you should perform a complete backup of the data and operating system before you begin your new installation. This way, you have a fallback plan if the installation fails or the server does not perform the way you anticipated.

When performing a new installation on a previously used server, you overwrite any data that was stored there. In this scenario, you will have to use your backup tape to restore any data that you want to preserve.

On the other hand, if you are going to upgrade an existing server, a known good backup will allow you to recover to your previous state if the upgrade does not go as planned.