The minimum and recommended hardware requirements for Windows 2000 Server have changed significantly compared to Windows NT Server 4, as shown in Table 8-3.

In addition to these items, I recommend that each computer have a Windows 2000 compatible network card. Otherwise, the computer will not be able to communicate and share resources with the rest of the network. An optional CD-ROM and floppy drive provide greater flexibility, especially for installation.

Windows 9x computers won’t be able to see an NTFS partition. If you are dual booting a Windows 9x/Windows 2000 machine, you may want to avoid using NTFS. Keep in mind that FAT and FAT32 partitions don’t offer the same level of security as NTFS.

There are two main ways to install Windows 2000 on your network. One way is to install directly from the CD-ROM on each computer, the other way is to copy the installation files to a network share and install over the network. You can only perform an over-the-network installation if the target computers are already on a working network and can read files from a shared network drive.

The Windows 2000 Server installation is divided into three phases: pre-copy, text mode, and GUI mode. You can make boot disks for non-Windows NT computers by running the makeboot program. For a Windows NT machine, use the makebt32 program. Both of these programs are located in the bootdisk directory of the Windows 2000 Server installation CD-ROM. You can type makebt32 a: to make a boot disk from a blank floppy in the A: drive.

Figure 8-1. Choose a licensing mode for Windows 2000 Server

After the computer reboots, you will log in normally by pressing the Ctrl-Alt-Del keys. Windows 2000 will try to detect any new hardware, and the Windows 2000 Configure Your Server dialog box appears. You can either configure the server now or choose to do it later.

To perform an unattended installation of Windows 2000 Server, you must use an answer file. This is a script that will automatically answer questions you would have had to manually respond to during the installation. Unattend.txt is an example of an answer file, which you can find on the Windows 2000 Server installation disk in the i386 directory. This file can be modified to suit your needs.

There are other resources available in compressed format in the file Deploy.cab, which is in the SupportTools directory of the installation disk. You can decompress the .cab file using Windows Explorer. After you decompress the file, you can read the Unattend.doc file for detailed descriptions of each of the answer file’s components, which are called keys.

A program called Setup Manager is also contained in the Deploy.cab file. This program can help you to create your own answer file. If you’d rather make the answer file manually, you can use a word processor and just modify the existing Unattend.txt file. Be sure you save the new answer file as a plain text document. The setup manager is described in detail in Part I.

Two tools you should be familiar with when performing unattended installations are syspart and sysprep. Sysprep, which is covered in Part I, is used to help create a disk image to be copied on similar, if not identical, computers. It’s best if you only use identical installation images on identical computers, otherwise you risk having incompatible settings copied.

Syspart is used when the computers aren’t similar enough to risk using sysprep. Syspart allows the file copy phase of the installation to be completed on one computer and then used by the other computers, even if the rest of the installation has to be customized. This can save a huge amount of time if you are deploying Windows 2000 to a large number of dissimilar computers.

If you plan to install Windows 2000 Server on multiple computers, you can create a template of the directory structure you want to be copied. All of the destination computers must already be connected to the network and be able to copy files from a shared network drive.

There can be as many as eight different distribution folder configurations used during the automated installation. These folders can include Windows 2000 system files and specific hardware device drivers, along with other types of files. A master copy of the distribution folders can be modified to quickly update new installations.

A service pack is an update to the operating system that usually fixes a reliability or security issue. Microsoft has made a few changes to the process of installing and maintaining service packs.

The first major change is that a service pack can be applied during the installation of Windows 2000. You can create a master installation directory on a network share that contains both the Windows 2000 operating system installation files and all available service packs. If you install Windows 2000 from this folder, all of the appropriate operating system and service pack files are installed automatically. Microsoft calls this process service-pack slipstreaming .

If Windows 2000 is already installed and you need to apply a service pack, run Update.exe /slip to install the necessary files from the service pack. You do not need to reapply the service pack if you’ve added or removed services from the system.

There are many reasons why Windows 2000 Server can fail to complete its installation successfully. Your best strategy is to take a systematic approach to figure out what went wrong. First, make sure your computer meets the minimum hardware requirements. Also, be sure you are using the latest disk controller drivers.

Many potential problems can occur while trying get the data from the installation media to the target computer’s hard drive. If you are going to perform a network installation, be sure that your target computer can connect to and read files from the shared source directory. A network failure or a lack of permissions will prevent you from accessing the installation files.

Even if you are installing from a local CD-ROM, a network failure, such as a broken wire or faulty hub, will prevent you from connecting to the domain controller. If you can’t connect to the domain controller, continue the installation as a member of a workgroup and attempt to connect after you finish the installation.

If you are installing directly from a local CD-ROM drive, be sure that your drive is supported by Windows 2000. If Windows 2000 doesn’t recognize the CD-ROM drive, you’ll have to replace it with a supported drive. You can find the most current list of supported hardware on Microsoft’s web site by searching for the Windows 2000 Hardware Compatibility List (HCL).

You may get an error message stating that a dependency service failed to start. This is probably caused by improper network settings. Be sure that the network adapter is compatible with Windows 2000 and that the installed protocols are supported on your network.

If the PDC is running WINS or DHCP services, you must stop them before beginning the installation of Windows 2000 Server. Failing to do so will prevent the installation program from converting those databases properly. You can stop both of these services using the Servicesapplet in the Control Panel.

To explain the process of upgrading to Windows 2000, we need to use a few terms that have to do with the Active Directory model. These terms and other Active Directory topics are covered in Section 8.4 of this chapter.

One of the great features of Windows 2000 is that, after you’ve upgraded the PDC, it will perform like a Windows 2000 Server to Windows 2000 clients, while simultaneously behaving like a Windows NT 4 PDC to other clients and servers. This means that, during the transition period, the network should continue to function normally and slowly improve as you add more Windows 2000 clients.

When upgrading the PDC, you’ll have the choice of creating a new domain or a child domain . There will also be a choice of creating a domain tree in an existing forest or creating a new forest . All of these options will be explained later.

The Security Accounts Manager (SAM) file needs to be upgraded to the new Windows 2000 format. You can store the Windows 2000 version of this file on any type of partition. There are two other files you’ll need to store during the upgrade: the log file and the system volume file ( SYSVOL). The log file can be stored on any type of partition, but the system volume file must be stored on an NTFS partition.

After you’ve finished upgrading the PDC, you should verify that the server is functioning properly before upgrading any other computers on the network. This is very important, because the new Windows 2000 domain controller’s configuration will be copied and used for all of the Windows NT backup domain controllers destined to become Windows 2000 domain controllers.

You should upgrade the backup domain controllers next. Even though you’ve thoroughly tested the new Windows 2000 domain controller and each BDC will use that configuration information, you should still test each upgraded BDC before moving on to the next. Upgrade them one at a time until all of the former BDCs are running well under Windows 2000. You can then move on to the member servers.

There are no special requirements when choosing which member server to upgrade first. You may want to verify that each is functioning properly before moving on to the next one.

After the PDC has been upgraded to Windows 2000, the network is running in mixed mode . This means that to Windows 2000 clients the domain controller will behave like a Windows 2000 domain controller, and to other clients it will behave like a Windows NT PDC. The main differences are that only Windows 2000 clients can take advantage of transitive trust relationships and group nesting, both of which are explained later in this chapter.

After the PDC and all of the BDCs are upgraded to Windows 2000 domain controllers, you can make a one-time only, one-way switch from mixed mode to native mode. After this is done, all clients can take advantage of transitive trust relationships and group nesting, both of which make access to resources more efficient. Prior to the switch, only Windows 2000 clients had access to these features.