Chapter 3 *

Installing Windows NT Server 4.0 *

Certification Objectives *

Planning the Hard Disk Configuration *

File System Considerations *

Per Server versus Per Seat Licensing *

Naming Conventions *

Preparing for Installation *

Exercise 3-:1 Preparing the hard disk before the installation process *

Exercise 3-2: Preparing the hard disk during the installation process *

Installation Methods *

Network Installation *

Exercise 3-3: Configuring Microsoft Network Client Administrator *

Exercise 3-4: Creating a network boot disk using the Windows NT Client Administrator *

Exercise 3-5: Creating a network share on a central Windows NT 4.0 file server *

Exercise 3-6: Connect to the network drive share that contains the distribution files *

Exercise 3-7: Starting a network share installation process *

Exercise 3-8: Network installation of the Windows NT 4.0 Server *

Exercise 3-9: Recreating Windows NT boot floppy disks *

Network Installation Switches *

Exercise 3-10: Using the text part of the Windows NT setup program *

Information Gathering *

Exercise 3-11: Using the NT Setup wizard—Information Gathering *

Installing Windows NT Networking *

Exercise 3-12: Using the NT Setup wizard—Installing Windows NT Networking *

Finishing Setup *

Exercise 3-13: Using the NT Setup wizard— Finishing Setup *

CD-ROM Installation *

Exercise 3-14: Starting a CD-ROM installation without floppy disks *

Installing NT on RISC-based Computers *

Upgrade Installation *

Unattended Installation *

General Setup Options *

Networking Options *

Advanced Options *

Exercise 3-15: Creating an unattended installation answer file *

Unattended Setup and the Uniqueness Database File *

Server Roles *

Primary Domain Controller (PDC) *

Backup Domain Controller (BDC) *

From the Classroom *

The Number One Problem in Installing BDC's *

Member Server *

Server Promotion Rules *

Removing Windows NT *

Deleting FAT Partitions *

Exercise 3-16: Deleting a FAT Partition *

Deleting NTFS Partitions *

Removing Windows NT from a FAT Partition *

Exercise 3-17: Deleting the system files to remove Windows NT from a FAT Partition *

Changing the Boot Loader in an Existing FAT Partition to MS-DOS *

Exercise 3-18: Changing the boot loader in an existing FAT partition to MS-DOS *

Exercise 3-19: Setting MS-DOS as your default operating system *

Troubleshooting the Installation Process *

Certification Summary *

Two-Minute Drill *

Self Test *

Chapter 3 Installing Windows NT Server 4.0

Certification Objectives

	Planning the Hard Disk Configuration
	Installation Methods
	Server Roles
	Removing Windows NT
	Troubleshooting the Installation Process

As discussed in the previous chapter, planning and preparation are an important part of the Windows NT Server installation process. In this chapter you will learn more about hardware requirements that need to be considered, the type of file system to be used, and type of licensing to choose. The bulk of the chapter describes two methods you can use to install Windows NT Server 4.0: across the network using a network share, or from a local CD-ROM. After a step-by-step description of the installation process, the chapter considers some special situations such as installation on a RISC computer, procedures for upgrading from a Windows NT 3.51 installation, and unattended installations.

As you complete the installation, you'll need to know what protocol the network will use for communication. Windows NT Server supports three protocol standards: TCP/IP, NWLink, and NetBEUI.

Another major section of this chapter deals with server roles, particularly the use of primary domain controllers (PDC's) and backup domain controllers (BDC'sfor domain management. The final sections describe methods for removing Windows NT from a server and troubleshooting an installation.

Planning the Hard Disk Configuration

Before you begin to install Windows NT 4.0, you must plan your hard disk partitioning scheme. Make sure to have a minimum amount of space available on the partition where you are installing Windows NT Server. You can set up a partition scheme prior to installation if you want to, or you can configure the partition scheme during installation. After the operating system is installed, system administrators can make changes to partitions using Disk Administrator—except that the system partition must be configured during the installation process, and cannot be changed without reinstalling.

Windows NT 4.0 Server requires a minimum amount of free space on the system partition (the partition where you install Windows NT Server). The absolute minimum recommended by Microsoft for Intel x86-type computers is 124 MB, but for satisfactory performance most networks require more space.

Two types of partitions are involved in Windows NT 4.0 installations: Thesystem partition and the boot partition. It is interesting (and sometimes confusing) that the "system" partition is where the operating system boots from, and the "boot" partition is where the system files are stored. These aren’t the only partitions available with Windows NT Server. Many other partitions can be set up for storing data and applications.

File System Considerations

You also need to decide how you want your file system configured. You can install Windows NT Server either with the traditional FAT (file allocation table) file system used for DOS, Windows 95, Windows 3.1 and Windows for Workgroups 3.11, or you can use NTFS, thenew technology file systemdeveloped specifically for Windows NT. Certain hardware dictates the choice of file type. For instance, you can install Windows NT on a RISC-based system, but you must use a FAT partition for the system files.

NTFS allows the administrator to set up up permissions that specify who can and cannot access files or directories (folders) on the drive. In other words, selecting NTFS allows the administrator to take full advantage of Windows NT's security features, and the FAT file system does not. If security is important in your corporate environment, NTFS is the preferred file system for Windows NT Server installations.

If you decide to use NTFS for your system partition, remember that the FAT file system is kept intact until the installation is completed. The setup program does not format a new partition to NTFS. The system partition is converted to NTFS on the first boot of the server after a successful installation. If you install Windows NT Server on a the FAT file partition, you have the option of converting the partition later on, using a CONVERT utility included with NT 4.0 Server that runs from a DOS session prompt.

You also need to considerwhether the computer you are installing will be a "multiboot" system. Generally a Windows NT Workstation is more likely to have a multiboot configuration than an NT Server is. Operating systems such as Windows 95 and DOS cannot access NTFS drives. If you need to share information between operating systems on a multiboot computer, FAT would be the appropriate file system choice. If you do install Windows NT as an NTFS partition, be sure that the system partition is set up as FAT so the other operating systems can boot properly. You can make a separate NTFS partition for your data by using Disk Administrator within Windows NT.

As mentioned earlier, if information security is essential in your company, then it's smart to take advantage of the NTFS security features. This also prevents someone from getting at the files on your server by using a DOS boot disk.

As you are studying for the Windows NT Server exam, it's a good idea to create a dual-boot installation, with Windows 95 in one partition and NT Server 4.0 in another partition. This will give you a hands-on experience with the installation process. Remember, Windows NT can be installed on an existing partition. For example, let’s say that you have a computer with Windows 95 already installed and want to install Windows NT Server. You can install to the same partition, but you cannot install Windows NT into the same directory as Windows 95! In addition, if you have multiple instances of NT Server or NT Workstation on one computer you must install them in separate subdirectories. The installation program will automatically update the boot loader menu.

If you have MS-DOS and Windows 3.1 or Windows for Workgroups, you can install NT Server in the same Windows subdirectory. This has the advantage that the automatic setup of icons and programs that already exist under Windows 3.1 or Windows for Workgroups. However, if you install Windows NT Server in the same subdirectory as Windows 3.1 or Windows for Workgroups the original operating system will no longer be available. With Windows 95 and other NT operating systems (4.0 or greater) you must reinstall all your applications so they are accessible from each operating system on your hard disk. if you don’t have access to a "live" network.

Most of the options mentioned above can be decided during the installation process. However, it is best to prepare and plan ahead of time. Planning your installation will save time and reduce the possibility of mistakes.

Per Server versus Per Seat Licensing

Another factor to consider is how your licensing will be configured. Licensing is important because it affects the cost of your network installation. Part of the planning process presented in Chapter 2 is network layout and design. The license arrangement is part of this design.

Windows NT comes with two licensing options. Per Server or Per Seat licensing.These options provide different types of client access.

With Per Server licensing, your license depends upon the number of concurrent connections. If you have 100 workstations accessing your server, you would need 100 client access licenses and a server set up with Per Server licensing. This is ideal for networks that only have one server because each workstation needs a separate client access license for each server it accesses. The server itself controls the number of connections. If you have Per Server licensing set up and a 100-client access license for Windows NT Server, workstation 101 will not be able to log on when all the licenses are in use.

With Per Seat licensing you can set up your server so that each client has an access license for as many servers as it can access. In this case, the server doesn’t control logons. Instead, you are licensed for the number of computers on the network, but the clients are not limited to one server. Each workstation can concurrently use multiple servers even though it only has one license.

During the installation process you must specify whether you want Per Server or Per Seat licensing. You can take advantage of the Licensing option in your Control Panel (see Figure 3-1) to tell Windows NT Server how many licenses you have. If you are unsure about which type of licensing is best for your organization, select Per Server. If Per Seat turns out to be what you need, you can make a one-time switch from Per Server licensing to Per Seat licensing.

Figure 1: Control Panel Licensing option

Exam Watch: You may see one or two questions about Per Seat and Per Server licensing. Be sure to keep in mind that you can make that one-time switch only from Per Server to Per Seat. You cannot switch in the other direction. Partitions and Fault Tolerance

When planning partitions—especially your system or boot partition—you should also consider fault tolerance and what methods you plan to use. Remember that mirroring, in which the data on one disk is exactly duplicated on a separate disk, is the only fault-tolerant scheme that can be used on the system or boot partitions. If you need more information about RAID (Redundant Array of Inexpensive Disk) techniques for fault tolerance, refer to Chapter 4 of this book.

To set up effective fault tolerance you need to keep data and fault tolerant information on a separation partition and/or disk.. A good rule of thumb is to set up one partition for the system files and then keep the remaining data on a separate partition. This enables you to set up the appropriate fault tolerant scheme that you choose.

Naming Conventions

When installing Windows NT 4.0 Server, it is important to understand the naming conventions. Microsoft uses a NetBIOS form of a UNC, or a Universal Naming Convention. Everything in Windows Networking relates to naming conventions. Each computer in the domain or workgroup is given a "friendly name." Windows NT converts the friendly name into the TCP/IP address, MAC address or other identifiable means of routing the information. The syntax for the UNC name is \\copmutername\sharename. Because the server name is determined during the installation process, make sure to give the server a name that makes sense. Often administrators will name the server something in relation to its responsibilities. The NetBIOS name of a computer can be 15 characters in length. For example, a Backup Domain Controller in the New York office for a company named Acme might be ‘AcmeNewYorkBDC’. Share names, on the other hand, can be up to 255 characters in length. An example of a share name might be NT40CDROM. In this case the UNC name would be \\AcmeNewYorkBDC\NT40CDROM. The slashes are important. Be sure to use the ‘\’ (backslash) and not the ‘/’ (forward slash). The server (or computer) name in Windows Networking should be preceded by two backslashes, with one backslash separating the server (computer) name and share name.

Preparing for Installation

In the following exercises, you will prepare your hard disk for Windows NT 4.0 Server installation. PLEASE BE WARNED that these exercises will delete the entire contents of your hard disk. Be sure to use a computer where you can clear the hard disk!

During the pre-installation process you can use FDISK and FORMAT from a DOS boot disk to partition and format the hard disk. For this exercise you will need both a DOS boot disk with FDISK and FORMAT as well as the installation CD and boot disks.

Exercise 3-:1 Preparing the hard disk before the installation process

For this first exercise, use the DOS boot disk that contains FDISK and FORMAT. Insert this disk into your floppy drive and boot your new computer. Before you begin the exercise, the hard disk should have no partitions on it.
When you get to an A:> prompt type FDISK

You will see a menu of options. Select Option 1, Create DOS Partition or Logical DOS Drive. (See Figure 3-2)

Microsoft Windows 95

Fixed Disk Setup Program

(C)Copyright Microsoft Corp. 1983 - 1995

FDISK Options

Current fixed disk drive: 1

Choose one of the following:

1. Create DOS partition or Logical DOS Drive

2. Set active partition

3. Delete partition or Logical DOS Drive

4. Display partition information

5. Change current fixed disk drive

Enter choice: [1]

Press Esc to exit FDISK

Figure 2: FDISK main menu

On the second menu select option 1, Create Primary DOS Partition.
The program asks if you want to use the entire available space on that drive to create the partition. Answer No. For this exercise, use 50% of the available space.
After the partition is created, leave the DOS boot disk in the drive and reboot.
This time when you see an A:> prompt, type FORMAT C:

You will be asked if you're sure you want to format that drive, as it will destroy all data. Answer Yes.
After a few minutes, the drive should format. You are now ready to start Windows NT setup and installation.

Alternatively you can partition the hard disk during the Windows NT 4.0 Server installation process. Exercise 3-2 shows you how to do that.

Exercise 3-2: Preparing the hard disk during the installation process

For this exercise, use the Windows NT Server boot disks and CD ROM. Insert the first boot disk into the floppy drive and boot the computer.
Insert the CD-ROM into your CD-ROM drive while disk 1 is booting up. A series of drivers loads and you will be prompted for the second disk.
Insert the second disk. When the option comes up, select Enter to install Windows NT Server.
The second screen shows what other devices were detected on the system. Your CD-ROM information should show up here. If not, you cannot continue installation unless you have the manufacturer's driver disk.
If you have the driver disk from the manufacturer and your CD is not showing in the list, press the S key to specify additional devices. From the device list, select the option that requires a disk from OEM manufacturer. Insert your disk and the program should load the appropriate driver into memory.
When you are satisfied with the choices in the additional devices list, press Enter to continue.
You should now see the License Agreement screen. Scroll down to the end, then press F8 to agree to the license information.
Next you see a summary screen containing information about your system configuration. You should see display information, keyboard information, pointing device, and so forth.
The next screen is the key screen in this exercise. This is where you select the partition where you want to install Windows NT. If partitions are already created (and of course if space permits), you can select one of them.
If you see free space without a partition type, you can select that free space, then select what type of format to apply. Your options in Windows NT are FAT and NTFS. For this exercise select FAT.
After a few minutes, the drive should format. You are now ready to start Windows NT setup and installation.

Installation Methods

When installing Windows NT Server 4.0 there are various methods one can use to configure the server. Windows NT Server comes with three boot disks and one CD. The installation files are no longer available on a floppy disks. The CD contains just over 600 MB of information, which would require more than 420 high density disks! Most new computers now come with a CD-ROM drive. If not, you can add one for less than $100.

You can install over the network or from the CD. The CD install can be done one of two ways.

	The Windows NT Server 4.0 CD is a bootable CD-ROM. If your server’s BIOS supports it, you can boot directly from the CD and start the installation process that way.
	If not, the three floppy disks that come with Windows NT Server are boot disks designed to allow users to start the installation process without a bootable CD-ROM drive.

The reason the "boot" process is so important is that the installation program runs on a low-level version of Windows NT. The installation can start in DOS, but certain related files load to allow the system to reboot and start the NT operating system.

The following sections first describe the network installation option, then examine the CD installation process.

Network Installation

In a network installation, files are made available (shared) from another network computer, such as another server or workstation. With a network boot disk, you access the shared directory to run the setup utility. Files are then transferred across the network to the local hard disk during the installation process. Figure 3-3 illustrates this process.

Figure 3: Network installation

This isn’t necessarily the best way to install across the network. Most network administrators would recommend copying the installation subdirectory (\i386) to the local computer and then executing the setup file to start installation.

You can use Windows NT Client Administrator to create a network boot disk for a target computer. Exercise 3-3 shows you how to configure Windows NT Client Administrator so it will install the client files and set up the share. You need to have the Windows NT Server CD in the CD-ROM drive to perform this exercise.

Exercise 3-3: Configuring Microsoft Network Client Administrator

From the Start menu, select Programs | Administrative Tools (Common).
Click the icon for Network Client Administrator.
When it opens, Network Client Administrator displays a screen showing four options (See Figure 3-4). The option Make Network Installation Startup Disk should be selected by default. Click OK.

Figure 4: Network Client Administrator

A screen displays (See Figure 3-5) that gives you the option to specify the path to the client files on the Windows NT CD. The destination directory where the files will be installed is also shown, as well as the name of the share.

Figure 5: Setting up Network Client Installation Files for client installs

The files on the CD are in the "clients" subdirectory. Most administrators use this path on their hard disk as well. The default share name is Clients.
Click OK. The program copies the Windows NT Client Administrator files from the CD to the hard disk and a configures a network share.

Now, in Exercise 3-4, you will create an Installation Startup Disk. This is sometimes confused with an Installation Disk Set, which is a complete set of disks used to install the network client software on a PC. The Installation Startup Disk is a single bootable disk that contains the network client software and allows you to access the network share and proceed with the installation.

To complete this exercise you will need at least one blank floppy disk. You should have also completed Exercise 3-3, which installs the client files and sets up the share on the central server.

Exercise 3-4: Creating a network boot disk using the Windows NT Client Administrator

Once the files are copied (in Exercise 3-3), you are prompted for what type of boot disk you want to make. The options are Network Client 3.0 for DOS and Windows or a Windows 95 boot disk (See Figure 3-6). For this exercise, select the Network Client 3.0 for DOS and Windows.

Figure 6: Target workstation configuration

Make sure you have the appropriate network card selected and that your floppy disk is in the drive. Then click OK.
The next screen (see Figure 3-7) allows you to specify the computer name, domain, and protocol in use on the network. This is also where you specify the default user name to use—a name that automatically displays in parentheses during the boot process of the disk. (If you are using TCP/IP, the TCP/IP setting will not be grayed out as in Figure 3-7. If you are not using DHCP—Dynamic Host Configuration Protocol, described in Chapter 2—you will have to specify the TCP/IP information.)

Figure 7: Network Startup Disk Configuration

When you have entered the settings you want to use, click OK. Network Client Administrator configures the disk as a boot disk that loads the network drivers, connects to the server, allows you to log on to the domain and allows you to access shared resources.

Once the boot disk is created, you'll go to the computer where you are installing the network operating system, and connect to the network. However, you must first set up a share on a central Windows NT 4.0 file server, as shown in Exercise 3-5, so the new computer can access the appropriate files.

For this exercise, you will need the Windows NT Server CD. At the file server, the administrator needs to share the CD-ROM so the installation files can be accessed from a remote computer. The administrator also needs a boot disk that loads network drivers on the new computer so it can connect to the shared resource.

Exercise 3-5: Creating a network share on a central Windows NT 4.0 file server

Make sure you are on a Windows NT 4.0 Server computer. You will need to log on as administrator in order to set up a share.
Double-click the My Computer icon
Find the drive letter that corresponds to your CD-ROM drive (in this exercise we use D:).

Right-click on the CD-ROM drive letter and you should see a pop-up menu, as shown in Figure 3-8.

Figure 8: Sharing your CD-ROM drive in Windows NT server

Select the Sharing option on the menu to open the Sharing window. Two options display: Not Shared and Shared As.
Select the option and enter the name of your share. For this exercise, type NT40CD.
Click OK. Record the computer name of the server and the share name you just created. You will need this information when installing the target computer with the network boot disk. (If you don’t know the computer name of the server, you can find it by opening the Control Panel and selecting the network option. Click the Network Properties icon and the General Tab. The computer name and domain name should be displayed there.)

When setting up the share on the server, some people suggest copying the entire contents of the CD to a directory on the hard disk and then sharing that, rather than using the CD. This can provide a faster response, due to the fact that hard disks have faster seek times than CD-ROM drives.

Once your network boot disk is created and you have a share set up on a central file server, you can use the floppy disk to access the share and then install Windows NT 4.0, as shown in the remaining exercises in this section.

Exercise 3-6 shows how to connect to the network drive share that contains the distribution files. For this exercise, you will need the network boot disk created in Exercise 3-4. You also need to have a share set up on a central server, as discussed in Exercise 3-5.

Exercise 3-6: Connect to the network drive share that contains the distribution files

Insert the network boot disk created in Exercise 3-4 into the floppy drive on the new computer.
As the computer boots, you should see the network drivers load and you should see a prompt that shows you a user name. You can press Enter to use the default user name or specify a new user name at this point.
After you enter your user name, you will have to supply a password just as if you were logging on to the server in the graphical user interface (GUI).
Upon proper connection and logon, you should see the message, "This command completed properly." This tells you that you are connected.
At the A:> prompt, type net use [drive letter]: \\servername\share name of Windows NT Server 4.0 CD. (Specify whatever drive letter you want; the net use command maps a drive letter to the share set up on the server.) Also make sure there is a space between the colon after the drive letter and your \\servername\share name.
This should connect a network drive to whatever drive letter you specified. Change to that drive and type dir.
You should see the directory of the Windows NT 4.0 Server CD

Once you have made your share on a central server and connected to that share via the network boot disk, you are ready to start the installation process on the target computer, as shown in Exercise 3-7. For this exercise, you should have completed Exercise 3-6, where you connect to a network drive to obtain the installation files.

Exercise 3-7: Starting a network share installation process

If you are on an Intel compter, you will want to copy the entire \i386 subdirectory from the CD prior to running the setup program.
Make a directory on your local drive called i386. Change back to your local drive and type md i386 from the root.
Then type copy [networkdriveletter:]\i386\*.* [localdriveletter:]\i386\*.* /s (The /s switch is an instruction to include all subdirectories).
After all the files are copied to your local i386 directory, change to your local i386 directory.
Type winnt /b at the command prompt. (The /b switch indicates the install will not use floppy disks.) This starts the installation process.

You can run the setup program across the network, but the recommended procedure is to copy the files locally, then run the install. The reason for this is to prevent excess activity on the network that will affect other users. Copying the files to your local directory is faster and it does not put as much strain on the file server.

To run setup from the file server, you would complete Exercise 3-6 first. Then, instead of copying files as in Exercise 3-7, you would change to the i386 directory on the network drive and then type winnt /b there.

Exercise 3-8 takes you through network installation, step by step. To perform this exercise you need to make sure, at the dedicated server, that the CD or directory containing the files is a shared resource. If necessary, refer to Exercises 3-5 to do this. Also make sure that you have a network boot disk that contains the MS-DOS Network Client 3.0; the disk should be set up with the appropriate network drivers for your new computer. If necessary, refer to Exercises 3-3 and 3-4 to complete this task.

Exercise 3-8: Network installation of the Windows NT 4.0 Server

Insert the disk into the floppy drive of the new computer and boot up the machine.
Refer to Exercise 3-6 to log on and set up a network drive connection to the shared directory on the server.
Next, you will want to copy the installation files to the local drive. Refer to Exercise 3-7 for information on how to copy the files and start the setup program.
Once the setup program starts, you should be prompted for the location of the Windows NT Server installation files. Confirm the default on the screen and press Enter.

There will be instances where as an administrator you may have to recreate the three installation boot disks. This can be done by using the Windows NT 4.0 CD. You can use a various switches in conjunction with the setup file, as shown in Table 3-1 later in this chapter. You can create the boot floppy disks from a working computer (either DOS, Windows 95, or Windows NT) with the Windows NT Server CD in the CD ROM drive. From the command prompt you would type Winnt /ox. (The /OX switch tells the winnt.exe file to only create the boot floppy disks and not to continue with the remaining part of the installation.)

Exercise 3-9 shows how to recreate the Windows NT boot floppy disks. For this exercise, you will need three blank, formatted high-density floppy disks.

Exercise 3-9: Recreating Windows NT boot floppy disks

On a working DOS, Windows 95, or Windows NT computer (either server or workstation), insert the Windows NT 4.0 Server CD into your CD-ROM drive
At the command prompt, type the driver letter of your CD-ROM drive. For example, type D: if your CD-ROM drive letter is D.
Change to the subdirectory of the computer type you will be installing. In this exercise type cd i386. (This is for an Intel-type computer.)
From the command prompt type winnt /ox

Follow the prompts on the screen. The setup program should prompt you for each of the three blank disks.

Notice that the setup program creates the disks in reverse order. The third disk is created first and the first disk is created last. The setup program tells you this, but many people overlook it. Be sure to label the disks accordingly.

Network Installation Switches

When you install Windows NT in a network installation setting, there are various switches that can be used (see Table 3-1). These switches can make a network installation go more smoothly and quickly. In most cases the ‘/b’ switch is used in a network installation. This switch controls whether or not floppy disks are needed during the installation. In most cases, an administrator would not want to use the floppy disks if they were installing over the network.

Switch	Result
/B	Floppyless install
/C	Tells setup not to check for free disk space on setup boot floppy disks
/E	Specifies command to execute when setup is finished
/I	Specifies the name of the setup information file
/F	Tells setup to not verify files as they are copied to setup boot floppy disks
/OX	Creates the boot floppy disks only. Doesn’t actually install Windows NT Server
/R or /RX	Specifies optional directory to be installed; the /RX denotes optional directory to be copied
/S	Lets the administrator specify the location of the source files. (for example, winnt /s:location or \\server\share\[path:]
/T	Tells setup where to put the temporary installation files. (for example, winnt /t:location). This must be a local drive
/U	This is used for Unattended Installations; requires /s
/UDF	Specify the Uniqueness Database File to use for unattended installation
/X	Do not created boot floppy disks

Table 3-1: Switches available for Winnt.exe or Winnt32.exe

Exam Watch: You may see one or two questions about the switch combinations on the exam. Be sure to look over Table 3-1 before you take the exam.

The setup program contains two parts. The installation process starts in text mode, and the second part uses the Windows NT GUI. The text part of the progam:

	Identifies correct hardware
	Confirms selection of partitions
	Confirms file system to be used for Windows NT
	Confirms the directory where Windows NT Server will be installed
	Copies essentials files to hard disk so setup can start NT upon reboot

Exercise 3-10 shows you how to use the text portion of the setup process.

Exercise 3-10: Using the text part of the Windows NT setup program

When you start the Windows NT Server setup process, the first screen shows the setup program loading various drivers.
The next screen identifies all hardware in your system. Your options are:

F1	Get help on installation (good for first-time installers)
Enter	Set up Windows NT Server
R	Repair a damaged Windows NT Server install
F3	Quit without installing

Press Enter to continue. You will see a list of drives that Windows NT setup detects. You can either press the S key if you need to specify additional CD-ROM drives, or SCSI adapters. (Note: If no CD-ROM drive is displayed, setup cannot continue unless you specify more devices by pressing S. You'll need the driver disks from your CD/SCSI manufacturer to install these devices.) Once you have installed the necessary devices, press Enter to continue to the next screen.
The next screen shows the License Agreement. Scroll down to the bottom of the agreement and press F8 to continue. F8 is the equivalent of agreeing to the License Agreement. If you don’t agree with the agreement, you can press Esc; in that case, setup will not continue.
On the next screen you should see a summary display of components on your system, including computer type, display, keyboard, keyboard layout, pointing device, and so forth. If everything looks okay, choose Accept. If something needs to be changed, use the arrow keys to scroll up to the option and make the change. When all the necessary changes have been made, press Enter.
The next screen allows you to specifying information about the various partitions. Be very careful on this screen: Pressing D will delete the highlighted partition! The F1 help on this section is good.
Once you select a partition, the setup program asks what you want to do with the partition. The options are: Format the partition (you can select from FAT or NTFS); Select a different partition; Convert a selected FAT partition to NTFS (if you are installing to an existing volume); Leave the current file system intact. This is a good choice, especially if you are installing for the first time.
The next screen asks you to specify the directory where you want to install Windows NT 4.0 Server
Next you have an opportunity to run exhaustive diagnostics on the drive you are installing to. If you press Enter the setup program will perform an exhaustive test of the hard disk. If you press Esc, it performs a short test of the drive and setup continues.
In the final screen of the text section, the program copies files to the specified path. After the files are copied, setup prompts you to remove floppy disk from disk drives and the CD from the CD-ROM drive.

Once all the temporary files have been copied to the appropriate subdirectories, the system reboots and the setup program continues to the graphical portion of the installation, where all the important configuration items are chosen. The graphical portion uses the Windows NT Setup wizard, which divides the installation into three sections. The first is called the Information Gathering, the second is Installing Windows NT Networking, and the final phase is Finishing Setup. Although the Windows Setup wizard proceeds directly through the three sections, we describe each section in a separate exercise for convenience.

Information Gathering

Exercise 3-11 guides you through the first section of the GUI setup, Information Gathering.

Exercise 3-11: Using the NT Setup wizard—Information Gathering

When the wizard displays, click Next to proceed. The first screen you see asks you to enter your name and organization. the NAME field is required, but the ORGANIZATION field is optional. Click Next to continue.
In the next screen you have to select how you want your server licensing set up. You have two options: Per Server or Per Seat. Refer to the discussion of licensing in the first section of this chapter if you need to review these options.
The next screen asks you to enter the CD key, which is a portion of the software serial number. The CD KEY, as on most Microsoft software, is located on the back of the CD sleeve or back of the CD jewel case. Enter the CD key and click Next to continue.
The next step is vital to the future of your server. This screen allows you to specify the name of the server computer. Enter the server name, then click Next to continue.
The next screen also contains an important decision-making step. You need to choose which server role you want your server to fulfill. Refer to the section on Server Roles near the end of this chapter if you need to examine these options.
Next, you need to enter the Administrator password. Be sure to keep this information in a secure location in case you should forget it! You will have to enter the password twice. The password must be 14 characters or less. Click Next to continue.
Next you must choose whether or not you want a repair disk. A repair disk is recommended for fixing crashed servers. If you don’t create a repair disk now, you can use the RDISK utility at a later time to create or update a repair disk. (Chapter 14 covers Windows NT repair disks.) Click Next to continue.
In the next screen, you need components for the final installation, including:

	Accessories: 13 components—calculator, screen savers
	Communications: 3 components—chat, Hyperterminal
	Games: 4 components—FreeCell, Solitaire
	Multimedia: 9 components—CD Player, sound schemes
	Windows Messaging: 3 components—Internet Mail

Choose the components you want to install, then click Next to proceed.

A transition screen shows that you are now entering the second section of the GUI setup: Installing Windows Networking.

Installing Windows NT Networking

Exercise 3-12 guides you through the second section, Installing Windows NT Networking.

Exercise 3-12: Using the NT Setup wizard—Installing Windows NT Networking

The first screen asks if your server will be wired directly to the network or be accessing the network remotely. In most cases your server will be wired directly to the network. Choose the appropriate option, then click Next to continue.
The next screen is the Install Microsoft Internet Information Server screen. This option is not needed unless your domain is wired directly to the Internet and you will use IIS to run your Web site. This can also be used for your company intranet, if you have one.
Next you need to select the type of network adapter to be used. You can choose Start Search to have the program look for an adapter card, or choose Select from list to identify the card in use. Once you have identified the appropriate card, click Next to continue.
Next you must tell Windows NT Server what protocols the network will use to communicate. Windows NT Server can communicate with the following protocols: TCP/IP, NWLink (IPX/SPX-compatible), and NETBeui (NetBIOS Enhanced User Interface). You can use the Select from List option to install some other protocols including DLC, PPTP, and Streams. When you finish this task, click Next to continue.
The next screen allows you to install Network Services such as DNS Server, Windows Internet Names Service, or Remote Access Server. Choose the services you want, then click Next to continue.
The next screen deals with the copying of network components. You may see one of several screens, depending on what protocols you selected previously.

	If you selected NWLink (IPX/SPX-compatible transport), you will be prompted for the frame type you want to use
	If you selected TCP/IP, you will see a question about DHCP as well as a

screen to specify the TCP/IP properties.

Answer the questions, then click Next to proceed.

Next you see a screen on network bindings. In most cases, the defaults are fine unless you need to remove or add something specific. Click Next to continue.
Now you have two choices. You can click Next to start the network and continue setup or you can click Back to stop the network if it is running. You have to choose Next to complete setup. If Windows NT Server is successful in starting the network, the setup program will continue to copy files. If starting the network is unsuccessful, you will be prompted to change to configuration of the network adapter—for instance, its IRQ (interrupt request identifier) or base port address.
The next screen varies according to the type of server you are installing. If you are installing a Primary Domain Controller, you simply have to specify the Computer Name and the Domain Name. If you are installing a Backup Domain Controller or a Stand Alone Server, you will have four fields to complete: Computer Name, Domain Name, Administrator User Name, and Administrator Password. You need the administrator's user name and password so the Backup Domain Controller can be added to the domain. If the username and password you specify do not have administrator privileges, the computer will not be added to the domain.
A transition screen shows that you will now enter the third phase of the graphical portion of setup: Finishing Setup.

Finishing Setup

Exercise 3-13 guides you through the final section of the wizard, Finishing Setup

Exercise 3-13: Using the NT Setup wizard— Finishing Setup

During this final phase you will see a series of screens that show the copying of files, set up of program groups, etc.
If you installed Internet Information Server 2.0, you will be prompted to verify the directory names for this application.
Next you will be asked to set up the time, time zone, and date information.
Next you will have an oppotunity to adjust your display properties. The setup program should check your system for the appropriate video adapter and install the driver. You may find that the setup program installs a standard VGA driver. If this occurs, all you have to do is change it after setup is completed.
The next screen you see depends on whether you chose to create a repair disk earlier in the setup process. If so, this is the place where the disk is created. If not, you can still create the repair disk at a later time with the RDISK utility.
Next you see the final copying of files for Windows NT Server.
Finally, Windows NT Server setup configures Windows Messaging and saves the final configuration. (Note: If you decide to install Windows NT Server on a NTFS partition you will see the format and installation take place on a FAT partition. After a successful installation Windows NT Server converts the partition to NTFS. The setup program ensures that everything was successful before it continues.) You should see a successful message, then the system reboots.

CD-ROM Installation

If the computer you are configuring has the option, you can boot directly from the Windows NT Server CD-ROM. The option to boot from the CD-ROM drive is configured from within the computer’s BIOS. If you are installing directly from the CD you should make sure that the CD-ROM drive is compatible with Windows NT. If the CD-ROM drive is not supported, you could run into problems later on if you try to configure Windows NT and the operating system prompts you for the CD. If your CD is unsupported under Windows NT Server 4.0, you should copy the contents of the appropriate subdirectory to your hard disk prior to installation and run the setup program from there.

Exercise 2-14 shows how to start a CD-ROM installation without floppy disks. For this exercise, you will need the Windows NT 4.0 Server CD.

Exercise 3-14: Starting a CD-ROM installation without floppy disks

Be sure that the computer you are installing has the CD-ROM drive set up as a bootable device. In order to work properly, this drive should be ahead of your hard disks in the boot order.
With the CD in the CD-ROM drive, reboot or turn on your computer.
Upon booting from the CD the installation program will start automatically.

Begin Q&A

I want to install Windows NT Server. I do not have a bootable CD-ROM drive and one of my boot disks is giving me a read error consistently.	On a working computer with a CD-ROM drive, use the winnt.exe setup utility to recreate the floppy disks by using the /OX switch.
I am getting ready to install Windows NT Server 4.0 on a computer with Windows 95 already on it. I want to keep Windows 95 available as a boot option and allow the Windows 95 operating system to view the contents on my NT 4.0 Server drives created with Disk Administrator What do I have to do?	In this scenario you need to use FAT for two reasons. The first reason is to enable Windows 95 to boot. If you install NT into an existing FAT partition, you want to retain the FAT partition so the original OS can still boot. Windows 95 can only read FAT partitions. If you were to create a new drive letter from within Windows NT Disk Administrator using NTFS, Windows 95 would not be able to read the contents of that drive.
I am installing across the network and do not want the setup program to even mention using the boot floppy disks. What command or switch do I need to use in this instance?	In this instance you would use the /b switch.
I want to create a network boot disk with Network Client Administrator (NCA). I’ve never run NCA before. In which subdirectory of the Windows NT Server 4.0 CD are the files kept? What is the default share name specified upon first-time setup?	\Clients and Clients
I am setting up a new computer and I want to name it NEWSALESCOMPUTER12A. I consistently get errors when I try to name the computer. Why?	NetBIOS names are limited to 15 characters.
I have only 80 MB of free space left on my main hard disk in my Intel computer and I want to install Windows NT Server 4.0. Can I install it?	You need at least 124 MB of free space to install Windows NT Server 4.0. It is recommended that you have more, but 124 MB is the bare minimum.

Installing NT on RISC-based Computers

One of the nice things about Windows NT Server is the fact that up can install it on multiple platforms. You can install Windows NT on a RISC-based system, but you must use a FAT partition for the system files. With RISC-based systems, the partitions must be set up ahead of time. Windows NT cannot set up and configure RISC partitions.

Upgrade Installation

If you currently have a server running Windows NT 3.51, you can install NT 4.0 Server as an upgrade. You can also upgrade from DOS, Windows 3.1, Windows for Workgroups 3.11, and Windows 95. One important note: Later releases of Windows 95 use the FAT32 file format, so make sure you are not trying to upgrade in a FAT32 environment. Windows NT 4.0 cannot access the FAT32 file system.

To upgrade from a previous version of Windows NT, you run the Winnt32.exe file. The first part of the upgrade process is the same as the normal installation. The setup program copies files from the CD into temporary directories on the hard disk (see Figure 3-9).

Figure 9: Windows NT Upgrade copies file to temporary directory on hard disk

Once the files are copied to the temporary directory, you will be prompted to reboot or stay in the current version of Windows NT (see Figure 3-10).

Figure 10: Rebooting during the upgrade

You will have the option to overwrite the existing operating system or install a dual-boot computer. If you choose to have a dual-boot machine, the boot loader menu will have a selection that allows you to choose the original operating system.

Of course, when upgrading you want to be sure and have a complete server backup available—in case the setup fails to complete and you have to revert to the original version. The following case study recounts an instance where a known good backup saved a company thousands!

A medical center had a Microsoft Windows NT 3.51 Server with various NT Workstations and Windows for Workgroups clients. This particular server was a Compaq ProLiant with a built-in CD-ROM drive and a Plextor 8 CD attached to it. An upgrade was scheduled for Saturday. Halfway through the installation, the server rebooted and the drive letters changed—which caused the entire upgrade to come to a screeching halt. Trying to reboot the server into Windows NT 3.51 did not work.

The only option was to reinstall Windows NT Server 3.51, then restore from backup. Then we started over on the upgrade. However, this time we disconnected the Plextor so that no drive letters would be changed in the middle of the installation process!

When upgrading from Windows NT 3.51 to Windows NT 4.0 you should be aware that the two products do not use the same default directory. By default, Windows NT 3.51 is installed in the \NT351 directory and Windows NT 4.0 uses the \WINNT directory. However, to maintain desktop, applications, and user information it is a good idea to put the upgrade files into the directory where the original version of NT had been located on that computer (in other words, in directory \NT351, not \WINNT). If you do not specify this directory when upgrading, you will end up with a dual-boot computer.

Of course the other big change with the transition from Windows NT 3.51 to Windows 4.0 is the GUI. The desktop and the product "look and feel" of Windows NT 4.0 is more like Windows 95. The change is similar to going from Windows 3.1 or Windows for Workgroups 3.11 to Windows 95. If your organization is used to the Windows 3.1 look of NT 3.51 it is essential to plan some training before the upgrade, to make end users comfortable with the new interface.

Unattended Installation

If you are deploying a large network with many servers and workstations, an unattended installation can make the operation much more efficient. Windows NT Server gives you the option to create an unattended answer file. Windows NT Server 4.0 CD contains a utility in the support\deptools\<computer type> subdirectory called SETUPMGR (see Figure 3-11). This is a GUI program that will help you set up answer files for unattended installation.

Figure 11: Setup Manager

There are three groups of options: General Setup, Network Setup, and Advanced Setup. The following sections describe how to set the options in each category.

General Setup Options

The General Setup Options screen covers basic information about the installation. The first tab is the User Information tab, shown in Figure 3-12, which includes User Name, Organization, Computer Name, and Product ID.

Figure 12: General Setup Options window in SETUPMGR

The General tab includes basic information about how setup will run in unattended mode. For example, you can specify that you want setup to confirm hardware during setup and you can list upgrade options. You can even specify that you want to run a program during setup.

On the Computer Role tab you choose one of the following as a role for the computer:

	Primary Domain Controller
	Backup Domain Controller
	Stand Alone Server in Domain
	Stand Alone Server in Workgroup
	Workstation in Domain
	Workstation in Workgroup

Along with the computer role you can specify the Workgroup name, if applicable, or the Account Name, if appropriate.

The License Mode lets you specify the license mode if the Computer Role tab indicates the computer is being set up as a Server or Domain Controller. If you have Workstation in a Workgroup or Domain you do not need to complete the License Mode tab.

The Time Zone tab is self-explanatory: it lets you select the time zone where the computer is located.

The Display Settings tab enables you to configure Display Properties during installation. You can set up refresh rate, horizontal and vertical resolution, bits per pixel, and various flags.

The Install Directory tab lets you specify where the installation files will be located. Choices here include: Use default directory, Prompt user for installation directory, and Specify installation directory now.

Networking Options

The Networking Options area includes information on network adapter card configuration, bindings, protocols, Internet, and modem. The General Tab, shown in Figure 3-13, allows you to specify whether you want a manual network installation or an unattended network installation. As part of the Unattended option, you can ask setup to detect and install the first network adapter it finds or you can ask it to detect specific adapters. Alternatively, you can bypass the detection stage and specify what adapters to install.

Figure 13: General tab of Networking Options window in SETUPMGR

On the Adapters tab, you can specify the network adapters to install.

On the Protocols tab, you can specify what protocols to load and install. Only the three basic Windows NT networking protocols are listed in this section: TCP/IP, NETBeui, and NWLink (IPX/SPX-Compatible Transport).

The Services tab allows you to specify which network services to install. The available options by default include, SNMP (Simple Network Management Protocol), Client Services for NetWare, and RAS (Remote Access Service).

The Internet tab is accessible only if the computer role in the General Setup section is set up as a server. If it is, this tab tells the setup program to install Internet Information Server (IIS) and gives configuration information for it. For example, you can tell setup where to install IIS and what subdirectories to set up for FTP, WWW, and Gopher. You can choose which options you want installed: Internet Service Manager (ISM) , HTMLA (an HTML version of ISM), and sample Web pages. You can also specify the guest account username and password to set up.

The Modem tab must be completed only if the Remote Access Service is selected in the Services tab. You can specify the COM port, modem type, manufacturer, provider, as well as the driver to install.

Advanced Options

The Advanced Options window allows advanced users to specify additional information for the unattended installation process. The General Tab in the Advanced Options, shown in Figure 3-14, lets you specify information on the Hardware Abstraction Layer (HAL), the keyboard layout, whether or not to reboot after text and graphic setup sections, and whether to skip the Welcome Wizard page and Administrator Password wizard.

Figure 14: General tab of Advanced Options window in SETUPMG

The File System tab allows you to select whether you want to leave the current file system intact or convert the file system to NTFS. You also have the option here to extend the partition beyond two gigabytes (a FAT limitation), if you convert to NTFS.

The Mass Storage tab allows you specify additional mass storage devices to install drivers for during installation.

The Display tab, Pointing Device tab, and the Keyboard tab all let you specify display drivers for the installation process.

The Advertisement tab lets you customize the banner text, logo, and background information for your installation. This is a spot where network administrators can gratify their egos by personalizing an unattended installation.

The Boot Files tab allows you to specify a list of boot files.

Once you’ve supplied the appropriate information, click the Save button in the main SETUPMGR window. You can give the file any name you want; it will have a .txt extension. There is also a New button on the main window, which clears all the information in the highlighted section so you can start over.

Exercise 3-15 describes how to create an unattended answer file.

Exercise 3-15: Creating an unattended installation answer file

To become familiar with the SETMGR, use the following scenario to create an answer file that will set up a server in a domain:

	Username: Bill Byttes
	Organization: Acme Inc.
	Computer Name: Bill’s Machine
	Be sure to confirm the hardware during setup
	Use the default install directory
	Configure the graphics device at logon
	Have setup choose the Bogota, Lima time zone
	Convert the existing file system to NTFS
	Add a 3COM Etherlink III Adapter
	You need to have TCP/IP and NWLink IPX/SPX loaded
	Be sure to skip the Welcome Wizard page
	Set up Per Server licensing with 100 licenses
	You have Novell workstations on your network, so be sure to add Client Services for Netware

Save the file, then try using it in an unattended installaiton. Remember to use the /u switch.

Once you’ve created the answer file, you have to tell setup where to get the file when the installation process begins. As mentioned earlier, the switch to use for an unattended answer file is: /u. For example you would type:

winnt /u:unattend.txt /s:{source}

Unattended Setup and the Uniqueness Database File

When deploying large networks, the unattended answer file is a great tool for speeding up the installation process. When you have ten or more machines to set up and install on your network, creating an answer file for each can be tedious. Windows NT allows the creation of what is called a Uniqueness Database File (UDF).

A UDF allows an administrator to specify information for each individual workstation or server that is being installed. Items such as computer name, IP address (in a static IP environment), time zone, and other information can be specified on a per-machine basis. To use a UDF, a valid answer file must already exist.

The UDF is created as a text file that contains information to be merged into the default answer file. This is done through the use of specific switches with winnt.exe or winnt32.exe.

An example of a command line use:

winnt /u:unattend.txt /s:d:\ /t:c: /udf:useridl,z:\udf.txt

Exam Watch: This information is not presented in depth in the exam. But be sure to understand the difference between the /U and /UDF switches and their respective purposes.

Server Roles

One of the most common concepts you will hear throughout your studies will be that of the domain. A domain, as shown in Figure 3-15, is a group of computers containing domain controllers that share account information and have one centralized accounts database. An administrator needs to determine what role the server or servers will play in the domain model. This is an important step in the installation process. A server can belong to only one domain.

Figure 15: Simple Domain Model showing various server roles

The Windows NT operating system has various roles a server can play. Remember, each server can only fulfill one role. The three possibilities are: Primary Domain Controller (PDC), a Backup Domain Controller (BDC), or Member Server (MS). Choosing whether the server will be a domain controller is a very important decision in the installation process. At the beginning of an installation, you must choose a role for the server you are setting up. If you make a mistake, you will have to restart the installation process.

Security is an important consideration when designing your network. An important role of domain controllers is to maintain security information—accounts information and the accounts database— for the domain. Domain controllers share a common security policy. The domain can also contain stand-alone (member) servers that do not maintain account and security information.

Primary Domain Controller (PDC)

A PDC is the central server in the network. There must be one—and only one—PDC per domain. If the server you are installing is the first computer in the domain, it will be installed as a PDC. If you are not installing a PDC, another PDC must already be connected to the network.

When installing the PDC you have the opportunity to name the domain. This is an important step, because it affects how the remaining servers and workstations will be named. When other servers and workstations are added to the domain, the domain name—the name you assigned when you set up the PDC—will become part of each server or workstation name.

The master accounts database resides on the PDC. Synchronization is the process of copying the security information (accounts and permissions) to the BDC's and performing periodic updates to keep the information up-to-date. This database is synchronized on the BDC's only. The domain administrator sets up synchronization and specifies how often it will occur.

Backup Domain Controller (BDC)

A BDC serves multiple purposes in a Windows NT Domain. The rule of thumb is to have at least one BDC per domain. As mentioned in the preceding paragraph, the BDC is responsible for helping maintain user account database information. When installing the BDC you must know the name of the domain to which you are assigning it. Whoever installs the BDC must have administrative privileges in the domain. Since BDC’s cannot be added to the network prior to installation, administrative privileges are needed to add machines to the network.

When a non-domain controller computer starts up, it goes through a process called discovery. That means the workstation looks across the network for a domain controller in its domain (and in all "trusted" domains; refer to the section "Trust Relationships" in Chapter 7). Once the workstation locates a domain controller, it uses that BDC or PDC for subsequent authentication.

BDC's can authenticate users when they log on to the domain. If the BDC fails to authenticate the user, then the logon request is passed to the PDC for authentication.

Another function of the BDC is to back up the PDC in the event of a PDC crash. Promotion of a BDC to a PDC by the domain administrator allows the BDC to take over the primary roll in the domain. This is the only type of server within Windows NT that can be changed without reinstalling.

From the Classroom

The Number One Problem in Installing BDC's

Once you have your planning done, and have made sure that your hardware is compatible (or that you have all of the OEM diskettes you might need to install any hardware that isn't on the distribution CD-ROM), NT installation is pretty straightforward.

Then you come to the dialog box that asks you to set the role for the server. You have three choices: PDC, BDC and member server. In the classroom, we frequently pretend that the students are servers, and set them up in domains. One student is the PDC, others are BDC's or member servers. We make a point of emphasizing that anyone who is a BDC must wait until the PCD is completely installed before they can perform their installation. In every class, there are several students who ask for help because they get hung up when they try to select a role. This is because the installation cannot find the PDC. Of course not—their domain partner has not yet completed the installation of the PDC!

The same problem arises at our clients' sites, where the environment is not so tightly controlled. There are two situations when it can occur. The first is when IP subnetting is place. The administrator is attempting to install a BDC on one subnet and the primary is on another subnet with no gateway (router) between them. Of course the BDC cannot find the PDC, and installation cannot continue. A more mysterious variation of this same theme occurs when the PDC is on one subnet and the BDC is on another and a router does exist between the subnets. The problem is solved, right? So why does the setup program return an error saying that the BCD cannot locate the PDC? If you find the answer to this question, please let us know!

The second situation occurs when the client is installing the BDC at a remote site and the PDC is on the net connected by a WAN link. Can you anticipate the problem? You guessed it! In our scenario, the administrator has planned a one-day trip to the remote site, many miles away, and arrives to find that the WAN link is down. Of course, the administrator discovers this during the installation. And the WAN provider promises the link should be up shortly—a story that's right up there with "the check is in the mail." For the next two days, the WAN link "bounces" up and down. If only you could predict the "up" times, you might be able to finish the installation! Too often, in this case, we get a request for support from the administrator who explains the symptoms to us, but hasn't ever checked to see whether the remote site can actually be reached from another system.

Remember the rule, no matter where you are: to successfully install a BDC, you must be able to contact the PDC. Period.

Member Server

A member server (sometimes referred to as a stand-alone server) is not a domain controller. Member servers have no responsibility in the accounts database and security information. They are used for application servers, print servers, or SQL servers. They are dedicated to process high-end searches, and non-security type functions. They help take the load off your domain controllers, which can concentrate on security and account information processing.

Server Promotion Rules

When something happens to the PDC, such as a crash or being taken offline for an upgrade, then the administrator must decide how the domain will function. If the PDC will be offline for an extended period, a BDC can be promoted to a PDC to take over the domain. When you try to bring a PDC back online, it will check the domain. If another BDC has been promoted to PDC, the original PDC will not be able to log on to the network. The administrator will have to use the Server Manager to demote the original PDC. (The Promote option on the Computer menu in Server Manager changes to a Demote option if the original PDC is highlighted on the list of computers in the domain.)

Stand-alone servers cannot be promoted or demoted. The only way to change the role of a member server is to reinstall Windows NT Server on the machine. The same applies to PDC’s and BDC’s. If an administrator wants to change an existing domain controller to a member server, the operating system must be reinstalled on that particular machine.

When promoting a BDC to a PDC there are two distinct possible scenarios. The BDC can be promoted while the existing PDC is still online or it can done after that the PDC has been taken offline.

If administrators know the current PDC will be offline, they can promote a BDC to a PDC to keep the domain alive. This is only recommended if the PDC will be down for an extended period. The master accounts database is synchronized on the BDC being promoted. When the BDC is promoted, the operating system automatically "demotes" the PDC, since there can’t be multiple PDC’s in the domain.

If a PDC is in danger of crashing, the domain administrator must decide how to maintain the domain. If the BDC is promoted while the PDC is already offline, the BDC uses the most recent update it has for the account database. If a domain synchronization hasn’t been done for a while and users have been added at the PDC, those users will be lost. When the original PDC returns to service, it will have to be demoted before it can return to the domain. This is done with the Demote option on the Computer menu of the Server Manager utility, shown in Figure 3-16. Remember, the user who performs this activity must have domain administrator permissions.

Figure 16: Server Manager with Promote to PDC highlighted

Exam Watch: Make sure you have a clear understanding of the promotion/demotion process. A Backup Domain Controller can be promoted to a Primary Domain Controller if the current PDC is going to be taken offline or if it has already been taken offline. If taken offline, the original PDC will have to be demoted before it can return to the domain. If the PDC is online when the BDC is promoted,the demotion takes place automatically. If the PDC is already offline when the BDC is promoted, the system administrator will have to user the Server Manager utility to demote the old PDC. Stand-alone or member servers cannot be changed to a domain controller without reinstalling Windows NT Server. Similarly, a domain controller cannot be converted to a member server without reinstalling.

Removing Windows NT

If you replace your server, you may want to remove Windows NT Server from the hard disk of the old server. There are a couple of ways to do this. You can delete the partition or you can simply delete the system files. The second method is better since it keeps other data intact

Deleting FAT Partitions

You can use FDISK from DOS or Windows 95 to delete a FAT partition. Some administrators think they can use Windows NT Disk Administrator to remove the system partition. This is incorrect. Disk Administrator lets you delete partitions, but not the system partition. Excercise 3-16 shows you how to delete a FAT partition.

Exercise 3-16: Deleting a FAT Partition

Boot your server with a DOS or Windows 95 boot disk that contains FDISK.EXE and FORMAT.COM
When you see the A:> prompt, type FDISK.
You will see menu that looks like Figure 3-17.

Microsoft Windows 95

Fixed Disk Setup Program

FDISK Options

Current fixed disk drive: 1

Choose one of the following:

1. Create DOS partition or Logical DOS Drive

2. Set active partition

3. Delete partition or Logical DOS Drive

4. Display partition information

5. Change current fixed disk drive

Enter choice: [1]

Press Esc to exit FDISK

Figure 3-17: FDISK Menu (Windows 95 version)

Choose option number 3: Delete partition or Logical DOS drive. You will see a list of partitions currently on your drive.
When you select a partition to remove, you will be prompted for the partition name, to make sure you selected the appropriate one.
Once the partition is removed you need to reboot so the changes will take effect. Keep the floppy disk in the drive and reboot to the boot machine
When you get to the A:> type FDISK to get back into FDISK. This time you will have to select option number 1 to create a new DOS partition to replace the old one. You will have to reboot one more time.
When you are back at the A:> prompt, type Format <driveletter>:

You will be prompted to be sure that you want to destroy all data on that drive.
After the drive is formatted it will be blank. You can now install a different operating system on it.

Deleting NTFS Partitions

You cannot use the FDISK utility from the DOS and Windows 95 operating systems to delete NTFS partitions because DOS and Window 95 do not recognize NTFS partitions. A utility such as DELPART, found in the Windows NT Resource Kit CD, can be used. DELPART is also available for download from the Internet. Use one of the major search engines, such as Yahoo or Excite, to locate it.

One other option for deleting an NTFS partition is to use the OS/2 version of FDISK. You can use the /D option to delete a partition.

If installing MS DOS 6 or later, you can go through the installation process to remove Windows NT and set up a new master boot record. Boot with the first disk of the installation set. When prompted, choose the Remove files option.

One final way to delete an NTFS partition is to initiate a new Windows NT Server installation and choose Delete partition in the text portion of the setup process.

Removing Windows NT from a FAT Partition

If you don’t want to destroy the data on your drive but want to remove Windows NT Server 4.0, you can delete certain system files. There are hidden, read-only system files on the drive root that you need to remove. You also need to remove some files in the Windows NT root (the subdirectory where you installed Windows NT Server—the default location is c:\WINNT). This procedure, described in Exercise 3-17, will work only if the partition is FAT, not NTFS.

Exercise 3-17: Deleting the system files to remove Windows NT from a FAT Partition

Boot to a DOS boot disk that has ATTRIB.EXE, SYS.COM, and DELTREE.EXE from DOS or Windows 95.
Once you are at the A:> prompt, type DELTREE c:\WINNT If WINNT is the Windows NT root (the directory where you installed Windows NT 4.0 Server).
There will probably be a file named pagefile.sys on the drive root. This file will also have to be deleted. Type DEL c:\pagefile.sys

There are a handful of other files you need to delete as well, but they are hidden system files. Type ATTRIB c:\<FILENAME> -h -r -s This will have to be done for the following files:

	NTLDR
	BOOT.INI
	BOOTSECT.DOS
	NTBOOTDD.SYS
	NTDETECT.COM

Once you have changed the attributes of the files listed above, then you can go ahead and delete them. Type DEL c:\<FILENAME> for each of the above files, to delete it.
The final step to uninstalling Windows NT Server is to restore the DOS Master Boot Record. Type SYS c: and reboot.
Upon rebooting, you should see a C:> prompt where you can load your new operating system.

Changing the Boot Loader in an Existing FAT Partition to MS-DOS

When Windows NT Server installs the master boot record on the drive is changed. In the event that a server needs to be setup with a different operating system, you need to make sure that the master boot record is rewritten for the operating system you have chosen. There a couple of ways to do this. The first is shown in Exercise 3-18, which shows you how to change the boot loader in an existing FAT partition to MS-DOS.

Exercise 3-18: Changing the boot loader in an existing FAT partition to MS-DOS

Use a DOS/Windows 95 boot disk and the SYS utility to configure the master boot record .
Insert the boot disk in the drive. Be sure that this disk has the file SYS.COM.
At the A:> prompt type SYS C:

Reboot the machine. Now you should boot to a C:> prompt.

If you do not want to remove Windows NT Server and simply want to make MS-DOS your default operating system, you can change the startup process within Control Panel. The procedure for doing this is described in Exercise 3-19.

Exercise 3-19: Setting MS-DOS as your default operating system

Select Start | Settings | Control Panel.
In Control Panel, select the System icon.
In the System properties window, select the Startup tab.
In this windows select MS-DOS in the startup box and enter the number 0 in the Show List for box.

I want to set up a server that will not be bothered with maintaining account information. What do I want to install?	You should set up a member server, which has no responsibility with the user account database.
My PDC just crashed and I need to keep the domain running. What do I do?	Use Server Manager to promote a BDC to PDC.
As a new member of Technical Support, I got a call from an Administrator who said that she was having a problem setting up a second PDC in her Domain Ajax. What should I tell her?	Only one PDC is allowed per domain.
I am the administrator of a domain that has one PDC and one BDC designated for each satellite office, but they are located in our local office. 15,000 client workstations are spread across the country in various satellite offices. (On the exam you will see questions that have network scenarios as big or bigger than this). Users complain that it takes a very long time to log on. The satellite offices are connected via T1 connections. What can I do to remedy this situation?	This question leans toward the Windows NT Server in the Enterprise exam, but it is still relevant here. An efficient way to speed logons in a WAN environment is locate a BDC in the remote office. Logging on across the WAN link can slow the process considerably.
I am the administrator of a Windows NT network that has one PDC and no BDC’s. The PDC has crashed and will not boot. The last known good backup is from yesterday. What can I do to bring the server back online?	Without a BDC to promote to PDC, your only choice is to reinstall Windows NT Server 4.0 and then restore from backup. Any changes between the backup and the server crash will be lost.
I am about to start the installation process with the boot disks. As I boot the machine, what do I have to type to get the installation process started?	Nothing. When you use the boot disks or CD-ROM drive for installations, the setup process starts automatically as the machine boots.
I have a machine whose CD-ROM drive is not compatible with Windows NT. What can I do to install Windows NT Server?	Copy the installation file to your hard disk and run the installation from there. Until you obtain an NT-compatible CD-ROM drive, you cannot install Windows NT Server from a CD-ROM.
I want to create an answer file for unattended installation. What utility is available to help me do this?	SETUPMGR on the Windows NT Server CD-ROM. You'll find it in the Support/Deptools/i386 subdirectory.

Troubleshooting the Installation Process

During installation you may have to do some troubleshooting if your installation doesn’t go smoothly. Here are some of the most common troubleshooting problems.

Faulty media is a common problem. The boot disks may have one bad disk or the CD itself may not be functioning correctly. There have also been many documented cases of insufficient disk space problems. These simple problems can cause financial headaches for companies, because it can take hours to track them down and resolve them.

Another item that can cause problems with configuration is an incompatible SCSI device, or a third-party drive like a Zip or Jaz drive. SCSI adapters devices (which is what most hard drives in servers are) can be one of the more challenging devices to get to work with Windows NT. The recommended method, of course, is to consult the latest edition of the Hardware Compatibility List (HCL). Microsoft Technical Support will not support you if you are having a problem with a device that is not on the HCL.

Another simple mistake—such as a wrong username, domain name or protocol—can make it seem that the machine you are configuring is not connecting. Oversights such as usernames and domain names, if not checked, can be mistaken for hardware problems. It’s much more economical to spend 15 minutes troubleshooting the basic network configuration than it is to buy a new NIC and toss the old one out.

Here's a real-life example:

A network engineer, when setting up a new PDC and various workstations, named the domain MainOffice1. In the course of setting up the network, one of the workstations was also named MainOffice1. This network is part of a WAN environment, with other domain controllers set up in trusted relationships. When finalizing all the machines and rebooting the PDC and workstations, the engineer encountered an error message. What do you think it was?

Right: duplicate names. This is a classic case of mere oversight. You always have to make sure that any and all names are unique. Whether it is a domain name or a computer name, it must be unique in the network.

Certification Summary

As part of the certification process you will need to be proficient in hands-on exercises; you'll also have to apply your knowledge to complex scenarios. The installation process for Windows NT Server accounts for about 10% of what you need to know on the exam and 40% of what you need to know in real life.

Be sure to know the hardware requirements for installing Windows NT Server 4.0. A good understanding of hard disk partitioning and planning is also important. Knowledge of basic NT administrative tasks, such as recreating the installation disks and the switches available for winnt.exe and winnt32.exe are also important. Be sure you have a good handle on the different security roles a server can have. You need to know how to maintain the domain environment, especially in the event of a PDC failure. When a Windows NT 4.0 Server installation fails, a good administrator can work through problems to get the server back online and restore proper functioning. Being able to troubleshoot the installation process is essential to the administration of your domain.

These items may be a small percentage of the entire exam, but they are invaluable for real-world situations. A good installation, with proper planning and preparation, can make domain administration a much easier process.

Two-Minute Drill

	Windows NT Server supports three protocol standards: TCP/IP, NWLink, and NetBEUI.
	The absolute minimum recommended by Microsoft for Intel x86-type computers is 124 MB.
	The syntax for the UNC name is \\copmutername\sharename.
	You can install Windows NT on a RISC-based system, but you must use a FAT partition for the system files.
	With RISC-based systems, the partitions must be set up ahead of time. Windows NT cannot set up and configure RISC partitions.
	Windows NT 4.0 cannot access the FAT32 file system.
	By default, Windows NT 3.51 is installed in the \NT351 directory and Windows NT 4.0 uses the \WINNT directory.
	The SETUPMGR utility in the support\deptools\<computer type> is a GUI program that will help you set up answer files for unattended installation.
	The Uniqueness Database File (UDF) allows an administrator to specify information for each individual workstation or server that is being installed.
	A PDC is the central server in the network. There must be one—and only one—PDC per domain.
	Member servers are used for application servers, print servers, or SQL servers. They are dedicated to process high-end searches, and non-security type functions.
	Stand-alone servers cannot be promoted or demoted. The only way to change the role of a member server is to reinstall Windows NT Server on the machine.

Review Questions: Just read, or Click HERE to launch interactive Self Test

What is the file to start the installation process that you type at the command prompt on a DOS machine?

setup.exe
install.exe
winnt.exe
winnt32.exe

You want to install Windows NT Server 4.0 and have all available security options accessible. What file system type do you want to choose during the installation?

HPFS
FAT32
NTFS
FAT

How many PDC's can a domain have?

unlimited
no more than 15
only 2
only 1

Your server name is SERVER and you share with the Windows NT 4.0 Server installation files is called NTCD. What do you type at the command prompt to connect to the network share when using a network boot disk?

net use g: \\SERVER\NTCD
net use g:=//SERVER/NTCD
net connect g:\\SERVER\NTCD
connect g:=\\SERVER\NTCD

You promote your BDC to a PDC while the PDC is off the network. What will happen you reinsert the old PDC into the network?

The administrator must first demote the old PDC to a BDC.
You can’t use it again, because only one PDC can exist within a domain.
You get error #45132: Duplicate computer name on network
None of the above

What switch do you use to recreate the setup boot floppy disks?

What method allows you to install across the network successfully on an MS-DOS workstation?

Use an Installation Startup Disk to connect to the Server with a share set up and run the setup.exe file to start the installation process.
Use a Network Boot disk to connect to the Server with a share set up and run the winnt32.exe file to start the installation process.
Use an Installation Startup Disk to connect to the Server with a share set up and run the ntsetup.exe file to start the installation process.
Use an Installation Startup Disk to connect to the Server with a share set up and run the winnt.exe file to start the installation process.

In which subdirectory on the Windows NT 4.0 Server CD-ROM are the installation files for an Intel machine located?

\intel
\386
\i386
\mips
\windowsnt

You have a BDC you want to make a member server. What do you have to do?

Reinstall Windows NT Server.
Demote the BDC to a member server.
Rename the BDC and change its server type in Server Manager.
Use the Server Setup option in Control Panel.

Which switch tells setup not to look for or create the floppy disks when starting setup?

You need to set up and configure three new BDC's for your WAN. What is the quickest and most efficient way to do this?

Boot each machine with the floppy disks and proceed with the installation with one machine one step behind the other.
Boot each machine with a separate CD and proceed with the installation with one machine one step behind the other.
Boot with a network boot disk and connect to a network share point with the installation files.
Boot with another server’s hard disk and copy the files over to the new drive in a master/slave relationship.

You are setting up a new Windows NT server on the network. This server will be a member server in the Finance department. You want to name this computer for the department where it is located. All servers, domain controllers, or non-domain controllers are set up and named in this fashion. The Finance department already has a BDC installed. What problems will you run into with this installation?

A. You cannot have a member server in the same domain as a domain controller.

B. Member servers must be installed prior to domain controllers.

C. The installation will result in a duplicate computer name on the network.

D. All servers must be set up at the same time before workstations can be added.

What is the default network protocol for Windows NT?

TCP/IP
NETBeui
NWLink
DLC

You have installed Windows NT Server 4.0 and now want to set up fault tolerance on your hard disk. Which of the following fault-tolerance methods can the system files be part of?

mirror set
stripe set without parity
stripe set with parity
volume set

A machine in your organization was formerly a Windows NT server. Now you find that you need this machine as a extra workstation. While logged on locally to the machine, you open the Disk Administrator and try to delete the main system partition. What will happen in this instance?

Disk Administrator will ask you to confirm your choice. When you click Yes, the partition information will be deleted.
Disk Administrator will ask you to confirm your choice, then show a message explaining the change that will take effect when the server is shut down.
Disk Administrator will ask you to confirm your choice, then shut down the server and remove the partition when you click Yes in the confirmation window.
Disk Administrator will return an error message saying that the operation cannot be performed.

When installing Windows NT Server, you selected Per Seat licensing. You find now that you should have chosen Per Server. What can you do in this instance?

Use the Convert.exe utility to convert the server from Per Seat to Per Server licensing.
Use the Control Panel Licensing option to change the licensing from Per Seat to Per Server.
You can’t change it unless you reinstall Windows NT Server 4.0.
You can change it, but you must contact Microsoft at licensing@microsoft.com.

When installing Windows NT Server you selected Per Server licensing. You find now that you should have chosen Per Seat. What can you do in this instance?

Use the Convert.exe utility to convert the server from Per Server to Per Seat licensing.
Use the Control Panel Licensing option to change the licensing from Per Server to Per Seat.
You can’t change it unless you reinstall Windows NT Server 4.0.
You can change it, but you must contact Microsoft at licensing@microsoft.com.

You are installing a multiboot system containing Windows 95, Windows NT Server and Windows NT Workstation. You want to be able to share files and data among the three operating systems. What file system should you choose?

NTFS
FAT
HPFS
FAT32

You are upgrading from Windows NT Server 3.51 that was installed with default settings. You want to maintain the icons, desktop settings, and basic Windows NT information. How would you accomplish this?

Use the winnt.exe utility and install into the WINT35 directory.
Use the winnt32.exe utility and install into the WINNT directory.
Install into the WINNT35 directory using the winnt32.exe file
Install into the WINNT directory using the winnt32.exe
None of the above

What Server Manager menu do you open to get the option to promote the BDC?

Computer
View
Options
Help

You are getting ready to start the installation process and decide to boot from the CD. When you put the CD in the CD-ROM drive and boot the machine, the computer still boots from the hard disk. What is a possible cause of this? (Select all that apply)

The computer you are setting does not have a bootable CD-ROM drive.
The Windows NT Server CD is not bootable.
The boot sequence on the machine is set to: Hard Drive, Floppy, and then CD-ROM.
The BIOS is not configured to include CD ROM boot as part of the regular boot process.

Where is the SETUPMGR utility installed by default?

SETUPMGR is not installed by default, but it is located on the Windows NT CD-ROM in the [CDDRIVE]:>SUPPORT\DEPLTOOLS\I386\
SETUPMGR is located in the Windows NT subdirectory on the hard disk after a default installation.
SETUPMGR is only available on certain versions of Windows NT prior to 4.0
SETUPMGR must be downloaded from www.microsoft.com\deployment\tools\i386\SETUPMGR.MGR

You’ve just installed a new server in a domain. You’ve verified that there are no evident errors on boot up. You check the domain name, and the name on the PDC matches the name on the server. You verify that you have a valid username and password. For some reason you cannot connect to the network. You check the Protocols tab in Network Properties on the new server and see the information shown in Figure 3-18. You check the same tab on the PDC and see the information shown in Figure 3-19. Why won’t the computers communicate?

Figure 18: Protocol setup on new server

Figure 19: Protocol setup on PDC

Each server must be part of its own respective domain.
There is a protocol mismatch between the two computers.
The frame type on the new server must match the IP address of the PDC.
You need to add the DLC protocol so the two computers will be able to communicate.

You are installing a BDC and encounter an error when the computer is being added to the network. Why is this happening?

The new BDC is using the wrong protocol.
The computer name is not valid for the domain.
The username/password you are using doesn't have permission to add computers to the domain.
The maximum number of BDC's in the domain has been exceeded.

You are installing Windows NT Server to be a member server. This machine is a dual-boot machine with Windows 95 OEM Service Release 2 already installed on it. You install Windows NT 4.0 Server in its own partition. You make sure to install the Windows NT partition as a FAT partition so Windows 95 can access the information on it. When you are in Windows NT you cannot access the drives that contain the Windows 95 information. You open Disk Administrator and see the screen shown in Figure 3-20. After reviewing the information in Disk Administrator, what conclusion do you reach?

Figure 20: Disk Administrator on a dual boot machine

The Windows 95 partitions were deleted and that is why they are showing Unknown.
The Windows 95 partitions are HPFS partitions which are not compatible with Windows NT 4.0 Server.
The Windows 95 partitions were formatted during the installation process and therefore show Unknown.
The Windows 95 partitions are FAT32 partitions, which are not compatible with Window NT 4.0 Server.

You have just installed Windows NT Server and realize you forgot to install Windows Messaging. Now you want to add it. What icon do you open in Control Panel (see Figure 3-21) to do this?

Figure 21: Control Panel

Network
Add/Remove Programs
Server
Internet

Which switch tells setup not to check for free disk space on floppy disks when creating them?

ANSWERS TO SELF TEST QUESTIONS

CThe Setup and Install programs do not exist in Windows NT Server 4.0. winnt32.exe is used to upgrade from previous versions of Windows NT Workstation and/or Server. DOS machines use winnt.exe
CHPFS and FAT32 are not compatible with Windows NT Server 4.0. Using FAT does not let you take advantage of all available security features built into Windows NT Server. The appropriate choice is NTFS
DThere can be only one PDC per domain and each PDC can only be master of one domain.
AIn answer B the switches are the wrong ones. In answers C and D the commands do not exist.
AThe administrator must demote the former PDC in Server Manager. If this is not done, the netlogon service will not run and the computer will not be able to participate on the network.
DThe switches in answers A, B, and C do not exist with winnt.exe
DThe ntsetup.exe and setup.exe files mentioned in answers A and C do not exist. In answer B, winnt32.exe is for upgrade purposes only. On a DOS machine you need to run winnt.exe.
CThe paths in answers A, B, and E do not exist. The path in answer D is for a non-Intel platform. The appropriate subdirectory is \i386.
AThe only change that can be made in a server role is promoting a BDC to a PDC. To change either a PDC or a BDC to a member server, you must reinstall Windows NT Server.
CThe /B switch tells setup to bypass and not create the floppy disks during installation. This is the switch used most commonly used during network installations.
CJust as there can be only one PDC per domain, so can each PDC only be master of one domain.
CYou will create duplicate computer names if you stick to the normal naming process for servers in this organization.
AAnswers B, C, and D are valid network protocols for use with NT, but they are not the default selection.
AAnswers B and D do not provide fault tolerance. Answer C is not an allowable fault-tolerant method of dealing with system files. However, system files can be placed in a mirror set (a RAID level 1 setting).
DBecause Windows NT Server 4.0 and Disk Administrator are running from the partition you are trying to delete, you will get an error message.
CWhen Per Seat is selected, you cannot change the licensing. Your only option is to reinstall Window NT Server 4.0.
BThe Licensing option in Control Panel lets you make a one-time change from Per Server to Per Seat licensing.
BWindows NT is not compatible with HPFS or FAT32 files, so that rules out answers C and D. Windows 95 cannot read NTFS, which rules out answer A.
Cwinnt32.exe is the correct installation file to use in this scenario, which rules out answer A. The default directory to Windows NT Server 3.51 is WINNT35, which eliminates answers B and D. Answer E therefore cannot be correct.
AThe Promote to Primary Domain Controller command is on the Computer menu. The other menus do not play a role in the promotion process.
A, C, & DChoice B is incorrect because the NT 4.0 Server CD is bootable.
ASETUPMGR was introduced with NT 4.0 Server but it is not installed by default. You could probably download it, but the URL supplied in answer D is fictitious. SETUPMGR is on the Windows NT CD-ROM, so that is the easiest way to obtain it.
BThe protocols that are loaded on two different computers must have at least one match and corresponding configuration information within the protocol (for example, subnet mask and frame type).
CThe username and password must have the ability to add computers to the network in order to install a BDC.
DWindows 95 is not compatible with HPFS, so answer B is not correct. The other two answers, A and C, are not correct because you installed Windows NT Server into its own partition.
BThe Add/Remove programs icon opens a window that has a tab for Windows NT Setup. You use this tab to add and remove basic components in Windows NT.
BThe /C switch tells setup not to check for disk space on floppy disks.