Free Tutorials
What is Internet
Internet Games
Learn TCP IP
Learn HTML
Learn CSS
Learn XML
Learn WML
Learn Access
Learn Data-VB
Learn Oracle
Learn SQL
Learn ActiveX
Learn C++
Learn CGI_Perl
Learn Interdev
Learn Java
Learn JavaScript
Learn Vbscript
Learn VisualBasic
Learn VC++
Operating systems
Learn RedHat
Learn Unix
Learn Winnt

Previous Page Main Page Next Page


Installing Windows NT Server

Setting up a Microsoft Windows NT server requires a lot of planning and decision-making. This chapter helps you plan for the optimum configuration for your server and guides you through the installation tasks required.

Having worked for a systems integrator, I've seen many types of Windows NT network configurations, used many types of hardware, and usually have run into a glitch here and there. In this chapter, I point out these pitfalls to prevent you from going through the same things I've had to work out on my own.

Examining the Hardware Requirements

Microsoft Windows NT Server is a very robust operating system, which is fairly intricate and requires a certain level of horsepower. It is scaleable, however, and takes advantage of any additional resources that can be offered for its use. Windows NT Server is limited only by the hardware chosen for the server, which is why it's a good idea to plan for the future.

Microsoft Windows NT Server is supported on different hardware platforms. The most popular of these platforms is the Intel x86 family of CPUs.

Although usable on a server with an Intel 486 CPU, almost any "server grade" PC has at least one Intel Pentium CPU. Because Microsoft Windows NT Server can take advantage of multiple processors, it is highly advised that the server you use for a Microsoft Windows NT Server has the capability of accepting additional processors in the future. Although your current needs might not indicate that multiple processors will be useful, down the road this might come in extremely handy, especially if there's a chance that other client/server software, such as Microsoft SQL Server, will be run on that piece of equipment.

Other hardware platforms supported for Microsoft Windows NT Server are the MIPS CPU, Digital Electronic Company's Alpha CPU, and the PowerPC chip. Although each of these platforms offers greater processing speed at the time of this writing, Intel's recently released Pentium Pro CPU, future offerings from Intel, and Intel-compatible CPU manufacturers may make the x86 platform the most cost-effective platform for you to use.

The memory requirements for Microsoft Windows NT Server can range from a minimum of 16MB of RAM to as much RAM as the hardware can accommodate.

Although it is possible to set up a Microsoft Windows NT Server with only 16MB of RAM, it is advisable to use at least 32MB. The 16MB configuration is sure to limit the speed of the server because it almost constantly swaps to disk, using virtual memory.

The formula for figuring out how much RAM is necessary for a Microsoft Windows NT Server is fairly simple. First, plan on 16MB of RAM just for the base product. This enables you to boot the server and offer basic file and print services. Again, I recommend starting out with 32MB; however, for this calculation, you can start with 16MB.

Next, determine the number of maximum concurrent connections to the server. In case you're not familiar with the concept of concurrent usage, this is the number of workstations that will have an active connection with the server at any given moment. Microsoft Windows NT Server requires server RAM for each of the open files that these clients have, so it is best to look at this in a worst-case scenario. Multiply the total size of open files that a user may be using from the server by the maximum number of concurrent connections to the server, and you will come up with a good starting point for figuring out how much RAM you need to add to the base configuration.

If your Microsoft Windows NT Server is supporting a group of 50 users, for example, but not all of these users will be connecting to the server at one given moment, do not use the number 50 (of course, if you can afford to allow for 50 users in this calculation, then do so!). Plan on 30 of those users maintaining a constant connection to the server, and that an additional 10 users may be on the server at any given moment. This brings your number of concurrent users to 40. If you can, assume that an average user will have 4MB of files open (I'm basing this on additional software that may be running on the server, such as Microsoft SQL Server or Microsoft Exchange Server). Now multiply the 4MB by 40 users, and you have an additional 160MB of RAM on top of the 16MB base configuration. This comes to a total of 176MB of RAM.

This scenario might be extreme, because users might not have files open on the server all the time. In fact, a lot of the installations I have performed didn't require any open files on the server because the users were only storing their own files on a server hard drive or using the print queues set up on the sever. In this case, there would be a minimal additional memory requirement beyond the base memory; however, the trend toward client/server applications has increased the requirements a great deal for a Microsoft Windows NT Server.

Again, if you decide that you only require the base memory, I strongly recommend not having less than 32MB of RAM on a Microsoft Windows NT Server.


The Microsoft Windows NT Server Performance Monitor is a great tool that you can use after you have gotten your server up and running to determine if additional memory is necessary. This is discussed in detail in Chapter 19, Performance Tuning and Optimization.

Besides the requirements to boot the Microsoft Windows NT Server and the memory requirements for users of client/server applications, you might be running software on the server that has its own memory requirements. An example of this is Microsoft SQL Server, which has a memory requirement based on the databases managed by SQL Server. Check with the software manufacturer first to find out what memory requirements exist for the server.


The more memory you can give a Microsoft Windows NT Server, the more freedom you will have to extend its capabilities. Whenever possible, it is best to take money saved in another area and buy extra memory for the server.

Because Microsoft Windows NT Server is shipped on CD-ROM, an NT-compatible CD-ROM drive is required for the server. Although there are workarounds to installing Microsoft Windows NT Server on a server without a CD-ROM drive, it is highly advisable to have this drive available because most applications for Microsoft Windows NT Server ship on CD-ROM as well.

Adequate hard drive space is also a concern when setting up a Microsoft Windows NT Server. When you are planning your server, you must think about the hard drive space required for these items:

  • The Microsoft Windows NT Server operating system
  • The individual users' home directories
  • Application software
  • Implementation of fault tolerance
  • Additional operating systems
  • Virtual memory

Early in the history of Microsoft Windows NT Server (when it was called Advanced Server), hard drives were still quite expensive. Now, however, adding a 4GB hard drive to a server is a fairly reasonable and cost-effective upgrade.

The Microsoft Windows NT Server operating itself requires approximately 100MB of space. If you are going to allow your users to store information on the server, you must allocate enough space by multiplying the number of megabytes you want them to be able to use by the number of users. Currently, Microsoft Windows NT Server does not enable you to limit the amount of space used by a single user, so you might need to overestimate the amount of space to be used.


You can limit the amount used by all users by creating a separate partition for user data. This limits the user's space to the size of the partition and alerts you to possible overuse by users if one user has a problem saving a document to this partition. Microsoft Systems Management Server or other third-party server management tools can alert you to low available disk space.

Some of the applications designed to run on Microsoft Windows NT Server have hard drive space requirements that can take you into requirements of many gigabytes. Microsoft's System Management Server (SMS) has requirements not only for the installation of the SMS software but, because it handles software distribution, it requires space for the "packages" of software it distributes. Besides that, it also requires Microsoft SQL Server software and a database running on SQL Server, so be sure to keep in mind the possible software that you might want to install on the Microsoft Windows NT Server.

Microsoft Windows NT Server requires the use of virtual memory in the form of a file called PAGEFILE.SYS. The space requirements for this file vary based on certain options set up in the configuration for your server, such as recovery. It's a safe assumption that 50MB to 100MB should be reserved for virtual memory.

Fault tolerance is a must for Microsoft Windows NT Servers. The exact amount of additional hard drive space depends on the type of fault tolerance implemented. If you decide that you require 4GB of usable hard drive space and you want to implement RAID 1 (mirroring or duplexing), you must plan for 8GB of hard drive capacity. The same 4GB of usable hard drive space, but with an implementation of RAID 5, only requires 6GB of total hard drive capacity. Although it is possible to expand the overall capacity of your server, planning up front and giving your server the necessary number of hard drives helps prevent you from having to bring down the server in the future to add hard drives.

Using the Hardware Compatibility List (HCL)

To assure compatibility between Microsoft Windows NT Server and pieces of hardware, Microsoft has wisely implemented the Hardware Compatibility List (HCL).

This list, which ships with the Microsoft Windows NT Server package and is updated on a regular basis and available via online services (such as CompuServe), lists hardware that has been tested and approved by Microsoft as Microsoft Windows NT Server-compatible.

The areas covered in the HCL follow:

  • x86 architecture uniprocessor computers
  • x86 architecture multiprocessor computers
  • MIPS RISC architecture computers
  • MIPS RISC multiprocessor architecture computers
  • Digital Alpha AXP RISC architecture computers
  • Digital Alpha AXP RISC multiprocessor architecture computers
  • Processor upgrade products
  • PCMCIA-tested hardware
  • SCSI host adapters
  • SCSI CD-ROM drives
  • Non-SCSI CD-ROM drives
  • SCSI tape drives
  • Other tape drives
  • SCSI removable media
  • SCSI scanners
  • Disk controllers
  • Hard drives
  • Wide SCSI
  • Storage cabinets
  • RAID systems
  • Video capture adapters
  • Video display support
  • Network adapters
  • Uninterruptible power supplies
  • Multimedia audio adapters
  • Modems
  • Hardware security hosts
  • ISDN adapters
  • Multi-port serial adapters
  • X.25 adapters
  • Third-party remote access servers
  • Keyboards
  • Pointing devices
  • Printers
  • PowerPC hardware

This is not a list of all hardware that works with Microsoft Windows NT Server; instead, it is a list of items that have been tested and given Microsoft's formal approval.


Use the HCL! If you call for support regarding a problem that might involve hardware, the technician might advise you to replace your current hardware with something listed in the HCL. Although this might not be the answer to your question, non-listed hardware often is cited as being responsible for symptoms of other problems.

It is important to note that although the HCL contains items that have been tested for use under Microsoft Windows NT Server, there is no clear indication of how some of these items interact with each other.

Although almost every Compaq computer is on the compatibility list and a lot of SCSI adapters are on the list, for example, this is not a guarantee that all these SCSI adapters function properly in almost every Compaq computer.


I'm only using Compaq as an example. Actually, I've found that Compaq computers are the most NT-compliant computers available.


It is your responsibility to make sure that the computer you are going to use as your server is of a server grade. In other words, even though an Epson Equity 4DX/33 is on the HCL, this doesn't necessarily mean that this computer gives you a great server. In fact, most hardware manufacturers will not offer support for one of their computers that was not designed to be used as a server if it is running as a server.

If you are acquiring equipment to run Microsoft Windows NT Server from a vendor, you should tell the vendor to make sure that the proposed hardware is included in Microsoft's Hardware Compatibility List.


Because new hardware is released faster than a new HCL is released, there may be newer, faster hardware available that you might want to include for your server. You might want to check with the hardware manufacturer to see whether its product has been approved for the next HCL or whether the product is undergoing testing for the next HCL.

Introducing Domains

The Microsoft Windows NT Server domain structure originated with Microsoft's LAN Manager back in the 1980s. It wasn't until Microsoft released Windows NT Advanced Server in 1992 that this structure was fully exploited.

The premise behind a domain is that you can have one or more servers that share a common element: the domain. The workstations on the network then can join the domain by logging into it (gaining access to domain resources, either with an authenticated logon using a user name and password, or by accessing domain resources using a Guest account).

The domain gives Administrators a single point of administrating user accounts, hard drives (known as shares), and network printers.

A Microsoft Windows NT Server network can include multiple domains that may or may not have a relationship with each other.

The servers that participate in Microsoft NT domains can be configured as domain controllers or as servers. The difference is that a domain controller has the capability to authenticate logons and can participate in other tasks that involve security, whereas a server is there purely to offer resources.

When implementing a domain structure, one server must be configured as the primary domain controller (PDC). This server will be the central repository of administrative information. Other domain controllers, known as backup domain controllers (BDCs), also will handle logon authentication and replicate administrative information between themselves and the primary domain controller.


When you are creating a new Microsoft Windows NT Server network, the first server you install should be the primary domain controller. You cannot create a backup domain controller without already establishing a PDC. You can create a server without having a PDC, but this scenario will not establish a Windows NT domain.

Computers running Microsoft Windows NT Workstation as their operating system can join an existing domain. Other clients running Windows 95, Windows for Workgroups, Windows 3.x, and MS-DOS also can participate in an existing domain.

Understanding the Windows NT Network Security Structure

The Microsoft Windows NT Server security structure is based on permissions; these give users the right to access a resource and specify the way in which the users can access the resource. Each Windows NT computer contains a security accounts database, known as the SAM. On computers running Windows NT Workstation, the SAM contains security information specific to that computer. On a Microsoft Windows NT Server Domain Controller computer, the SAM contains security information about the local machine and the entire domain.

User Accounts

Every user account in the domain has a unique ID called the security ID (SID). The SID is the primary way in which NT tracks permissions. These permissions are placed in an Access Control List (ACL).

At the time of logon, each user is assigned a security access token, which includes the user's SID and information on group memberships and the associated SIDs for those groups. The security access token is created by Windows NT, and a copy is passed to whatever process the user requests to access. Validation of permission to perform that process[md]whether the process is running a program or just accessing a file[md]is based on the security access token interacting with the ACL.

A user account (UA) is the external identification used for clients of the Windows NT domain that want to have an authenticated logon to the domain (the SID is used only internally and is never seen by users or Administrators). The user account includes information about the client, such as the user name (the ID used for logging onto a Windows NT network); permissions; and, among other administrative items, rules, which are known as profiles.

When a user account is deleted from a domain, the SID associated with that account is never reused. Even if the same user name is used with a new account, a new SID is generated.

By default, an account named Guest is established on a Microsoft Windows NT Server. This account is given the default permissions that have been established for the domain in which the server participates. The Guest account must be enabled if you want to use it.

Other information that is part of a user account includes a full name, a free-form description, a password for logging into the domain, domain group memberships, password restrictions, the location of a home directory, a logon script, and the range of hours allowed for a permissible logon, among other items. These items are discussed in depth in Chapter 15, Administering the Server.


Microsoft Windows NT Server domain groups are containers that can logically group together multiple user accounts. This way, permissions can be assigned to a group, and then any user account placed in that group is assigned those permissions. Each group has a SID associated with it, and that SID is included in a user's security access token.

There are two types of Microsoft Windows NT Server domain groups: local groups and global groups.

A local group is the type of group that a single domain NT network would use to group user accounts to assign permissions. A global group is a way to define a group of user accounts that then can be added to local groups that exist in other domains.

Microsoft Windows NT Server comes with predefined groups that already have permissions assigned based on the permission (or lack of) to perform administrative tasks. The Administrators group, which already includes the default user account, Administrator, gives permission to all system functions. The Backup Operators group only has access to the Windows NT Backup applet, it can log onto the server at the server console, and it can shut down the server. The Domain Users group only allows access to the server from a remote computer (not from the server console itself).

The User Manager for Domains enables Administrators to add user accounts to groups, create user accounts, and create groups.


As mentioned earlier, a Microsoft Windows NT Server domain consists of one or more servers and clients. This logical grouping of computers is an efficient way to administer resources.

A Windows NT network can include more than one domain. To administrate more than one domain from the Server Manager applet, however, the Administrator must have an account on all the domains he or she needs to administer.

Trust Relationships

To allow one domain to access resources on another domain, you must establish a trust relationship. Trust relationships also can allow for centralized administration of networks that go beyond a single domain.

Trusts are a one-way relationship, although you can create two trusts that have two domains that trust each other.

Planning Your Installation


In this section, I refer to the Microsoft Windows NT Server CD-ROM's \I386 directory. This assumes that you are dealing with an x86-based server. If you are dealing with a MIPS, Alpha, or PowerPC, then the proper directories are \MIPS, \ALPHA, and \PPC, respectively.

You must make a few decisions before you can install Microsoft Windows NT Server.

First, you need to determine how you are going to install Microsoft Windows NT Server. The preferred method is by booting from the Microsoft Windows NT Server Disk One and then following instructions. You can use this method if you have a CD-ROM drive supported on NT.

Another method is the command-line installation from MS-DOS. The executable file that enables you to install from DOS is WINNT.EXE. Usually, you use this method only if you do not have a CD-ROM drive supported on NT or if you do not have any CD-ROM drive at all and want to perform the installation from a network drive where the Microsoft Windows NT Server files are stored.

The proper syntax for the WINNT.EXE command follows:

WINNT [/S[:]sourcepath] [/T[:]tempdrive] [/I[:]inffile] [/O[X]] [/X | [/F]
[/C]] [/B]

Where /S[:]sourcepath specifies the source location of Windows NT files. This must be a full path of the form x:\[path] or \\server\share[\path]. The default is the current directory. And /T[:]tempdrive specifies a drive to contain temporary setup files. If a drive is not specified, Setup attempts to locate a drive for you. If you have developed a custom install script, you should use the /I[:]inffile parameter, which specifies the filename (no path) of the setup information file. The default is DOSNET.INF.

Other switches that you can use with WINNT.EXE follow:

/B Specifies floppyless operation.
/C Skips the free-space check on the Setup boot floppy disks you provide.
/F Does not verify files as they are copied to the Setup boot floppy disks.
/O Creates boot floppy disks only.
/OX Creates boot floppy disks for CD-ROM or floppy-disk-based installation.
/X Does not create the Setup boot floppy disks.

Using the WINNT.EXE installation program uses double the disk space for the Microsoft Windows NT Server installation process because it copies the contents of the \I386 directory on the CD-ROM to the local hard drive. After the installation is complete, the installation files are deleted.

A Microsoft Windows NT Server 3.51 installation or a PC running Microsoft Windows NT Workstation can execute the file WINNT32.EXE, which is the 32-bit version of the method described here.


Some people just copy the \I386 directory from the Microsoft Windows NT Server CD-ROM to an \I386 directory on the server's hard drive and then run WINNT /B just to avoid dealing with floppy disks. This method also enables you to make on-the-fly changes to your Microsoft Windows NT Server configuration after installation without having to find the original Microsoft Windows NT Server CD-ROM. The space this \I386 directory uses is minimal[md]only 40MB.

You also need to know where you are going to install Microsoft Windows NT Server (which hard drive, an existing partition, a new partition, and so on). You need to decide whether you will upgrade an existing copy of Windows 3.x by choosing to install to the same directory in which Windows 3.x exists, or whether to create a new directory for Microsoft Windows NT Server 4.0.

Unless you are upgrading an existing Microsoft Windows NT Server 3.51 server, I recommend that you perform a clean installation of Microsoft Windows NT Server 4.0 by first formatting your hard drive and then booting from the first Microsoft Windows NT Server 4.0 disk. This is my preference because it ensures that all files on the hard drive belong to this version of Microsoft Windows NT Server. Also, if this server will be a dedicated NT server, there really is no reason to have any other operating system on the server.

If there is information on the server's hard drive that you want to retain, you can leave it on, but if the installation program wants to upgrade an existing Windows installation, choose to install to a separate directory.

You need to consider certain parameters that will have a big impact on your network before you install Microsoft Windows NT Server.

You must come up with a name for your NT domain. If you already have an NT network and will be adding this server to the existing domain, you won't have to think about this (just know the name of the domain to which this server will be added).

The Microsoft Windows NT Server installation program asks whether the installation partition should be formatted as FAT or as NTFS. The decision to format the partition as FAT (or to leave the partition formatted as FAT without affecting existing data) can be changed in the future; you can convert FAT to NTFS, but you cannot convert NTFS to FAT, so choose NTFS only if you are sure that is the file system you want to use.

Under normal circumstances, use the NTFS format. FAT has a limited amount of security compared to the security offerings for the NTFS file system. To find out more about NTFS, check out Chapter 6, Windows NT File System Management.

A name for the server also is needed for the installation routine. Depending on the scope of your network or your domain, you should consider making this name descriptive enough to be able to pick out the server from a list of servers.

Perhaps one of the most important decisions you need to make prior to installation is defining the role of the server.

Microsoft Windows NT Server gives you three choices for the role of the server:

  • Primary domain controller (PDC). Usually the first server in a domain is designated as the PDC. Each domain must have a PDC and cannot have more than one PDC. A PDC handles logon authentication and holds the security accounts database (SAM).
  • Backup domain controller (BDC). You can choose this option only if there is already a PDC running on the network. The BDC also authenticates logon requests, handles administrative tasks, and keeps a copy of the SAM synchronized with the copy of the PDC. In the case of a failure on the PDC, a BDC can be promoted to a PDC.
  • Stand-Alone server. A stand-alone server does not handle logon authentication. Usually, you designate a server as a server when you require the resources of a server to handle specific tasks (such as running Microsoft SQL Server) and do not want that server to also have the overhead of administrative tasks.


If you are installing your first server, make it a primary domain controller (PDC). Without a PDC, you cannot manage domain user accounts. The NT installation program enables you to choose server or the server role even if there is no PDC, but it does not enable you to create a backup domain controller without seeing an existing PDC. Implementing a server without having a PDC is almost the same as having a PC on the network running Microsoft Windows NT Workstation, which also lacks domain administration tools.

Implementing Domains

When planning a Microsoft Windows NT Server domain, you have to look at the big picture. Not only should you evaluate your current networking structure (or proposed structure), but you should think ahead as well, because planning ahead is the best way to help ensure that you will not have to plan your structure from scratch again later.

You can choose from four domain structures:

  • Single domain model: Places one or more Microsoft Windows NT Servers into one domain. Because all user accounts and all resources are within the same domain, there is no need to set up trust relationships, and all the administrative tasks can be handled in one place. This is Microsoft's preferred model for networks with fewer than 15,000 users. If you have different departments that need to administrate their own resources, you might want to pick one of the other domain models.
  • Master domain model: This is a good choice if your organization does not require breaking up clients into separate domains, and you want to administer all the user accounts from one domain. This model differs from the single domain model because other domains are created as resource domains. These resource domains do not include user accounts, but instead are trusted domains that offer users in the master domain access to disk space, applications, printers, or any other type of shared resource. Usually this scenario has multiple departments, which all are administrated through the master domain, but each department can have its own set of resources.
    If you are starting an NT network that will involve only one department or the whole company and you do not see the need to create multiple domains, then plan this first domain as your master domain. The master domain structure puts all your user accounts into one domain. You might want to implement resource domains in the future. A resource domain does not have any user accounts, so it does not handle any logon authentication. It does have resources, such as disk space or printers, however, within the domain. One resource domain per department or workgroup and all the user accounts in its own domain make up the master domain model.
  • Multiple master domain model: This choice provides more than one account domain and one or more resource domains. This allows independent administration of different groups of user accounts, but trust between the resource domains and the account domains allows resources to be used by any of the users, regardless of the account domain to which they belong.
  • Complete trust domain model: This choice has multiple domains that all trust each other. Like the multiple master domain model, administration of the user accounts is broken up into different domains. This domain structure may not involve separate account and resource domains, however. The complete trust, which implies that each domain trusts all the other domains, shares its resources.

Domains Versus Workgroups

An alternative to defining a domain is using a function of Microsoft networking, which is known as workgroups.

Any computer that is running the Microsoft networking client software, but is not defined as part of a domain is automatically part of a workgroup. If you implement an NT server as a Stand-Alone Server, it is then part of a workgroup.

A workgroup can include one or more computers. Computers that do not participate in a domain, and are therefore part of a workgroup and are responsible for their own security and administration.

Workgroup computing is a good alternative for a small number of computers that do not want to utilize centralized administration, but do want to include a computer running Microsoft Windows NT Server because of some of the services that an NT server can offer, such as Dynamic Host Configuration Protocol (DHCP), or Remote Access Services (RAS).

An Installation Overview

After you complete the installation process for Microsoft Windows NT Server 4.0, you should have a working server that you then can tailor by adding user accounts, setting permissions, creating shares, creating print queues, and performing many other tasks.

Keep in mind that your first task, the installation, should result in a bootable, working Microsoft Windows NT Server. Certain configuring should be saved until after the installation process, once the server is up and running. For example, there is no need to define any video mode beyond VGA, unless the installation program automatically detects it. An incorrect video driver could cause the computer to freeze during an installation process, which could force you to start the installation program all over again. Sound support or any other peripheral device beyond the network adapter should also be saved for after the installation process has completed and the server has been restarted.

Make sure that you have read the section "Planning Your Installation" earlier in this chapter. Have a disk handy to make your Emergency Repair disk, and try to be sure that your hardware, including your network interface card, is functioning properly.

The first portion of the installation, which is the character-mode portion, prepares your hard drive's master boot record by copying the necessary Microsoft Windows NT Server system files.

The second part of the installation, which is the graphical-mode part, enables you to configure basic networking and sets up the parameters for the server whenever someone is logging on locally at the system console.

A properly planned Microsoft Windows NT Server installation should not take more than one hour.

Installing Windows NT, Step by Step

The first thing you need to determine before installing Microsoft Windows NT Server is whether you want to have another operating system on the server in addition to Windows NT.

If you are working with a system that already has an operating system on the hard drive, you might want to use the FDISK or FORMAT command to start with a clean hard drive. Of course, you will want to use these commands only if you will be using the server as a dedicated Microsoft Windows NT Server. If you also need to be able to boot MS-DOS, Windows 95, or OS/2 on the server, you will be able to do that by using NT's boot-loader[md]a menu that is displayed each time a PC runs Windows NT, which enables you to choose an operating system from which to boot.

As you learned earlier in the section "Planning Your Installation," you can use different methods for installing Microsoft Windows NT Server. This section assumes that you are using the preferred method, which is booting from the Microsoft Windows NT Server Setup Boot Disk and installing from the Microsoft Windows NT Server CD-ROM.

Booting the Server

Insert the Microsoft Windows NT Server Setup Boot Disk into your drive A, which must be a 3 1/2-inch drive. After the system starts to boot from that disk, you see a message that says Setup is inspecting your computer's hardware. Next, the blue screen present throughout the character-based portion of Setup appears, and messages across the bottom of the screen inform you that the Windows NT Executive and the Hardware Abstraction Layer (HAL) is loading. You then see a prompt to insert Setup Disk 2 into your drive A. Press Enter to continue the installation.

More drivers then are loaded, which give the installation program support for hardware and file systems. After these drivers are loaded, you see the screen font change to a smaller font and the actual booting of Windows NT occurs. The first line identifies the operating system by name, version, and build number. The second line tells you the number of processors Windows NT can see, the amount of physical memory detected, and always shows that the multi-processor kernel is loading. If you are running a single-processor server, the next time Windows NT boots, the single-processor kernel loads, but during installation, the multi-processor kernel always loads.

The next screen gives you the following choices on how to proceed with your installation:

  • You can get more information about the Microsoft Windows NT Server installation process by pressing the F1 key.
  • You can proceed with the installation by pressing Enter.
  • You can repair a previously installed copy of Microsoft Windows NT Server that may have been damaged.
    This is the option you should choose if you are unable to boot Microsoft Windows NT Server or if you have made changes to your configuration and want to reinitialize the installation. You then can press R to be prompted to insert an Emergency Repair disk. You can build the Emergency Repair disk during this installation process, or you can create it at any time from the command line.
  • You can quit the installation process and restart the server.

Press Enter to continue with your installation.

Specifying SCSI and IDE Controllers

The next screen informs you about Windows NT's detection of SCSI and IDE controller chips. In order to have a bootable Microsoft Windows NT Server system, you must have NT install support for your boot device. You also need to have NT install support for the controller to which your CD-ROM drive is attached (of course, this might be the same as the controller for your boot device, but if they are different, you must ensure that support for both is installed).

You have the option of letting the installation process try to identify which devices are in your system (this process is known as auto-detection). If your controllers are identified properly through this process, you can be assured that support for your boot drive and CD-ROM will exist the next time the Microsoft Windows NT Server installation routine reboots your server. If you have purchased a controller that came with its own set of Windows NT drivers or you know that the installation routine will not recognize your controller and you have a disk to use that contains the appropriate drivers, press the S key.

If you press the S key, the installation routine tells you that it did not detect any devices, and you then can press the S key again to choose from a list of drivers that come with Microsoft Windows NT Server. The last item on that list is Other, which enables you to point the installation program to a floppy drive.


You might want to try out auto-detection just to see whether Windows NT's native support includes support for your controllers. You should go straight to the S routine if your drivers are newer than the release date for Microsoft Windows NT Server 4.0.

Another reason to choose the S option is if you attempt auto-detection and your server freezes up. It is possible for a server to seize during the auto-detection stage yet still be able to function properly through the entire installation process if auto-detection is skipped.


If you use auto-detection and no devices are found, you can choose from the list of devices that ships with Microsoft Windows NT Server. Chances are that choosing from this list after trying auto-detection won't do you any good, however, because if you had a compatible device, auto-detection would have picked up on it.

If you do choose to use auto-detection, the installation program prompts you to insert Setup Disk 3 into the floppy drive. After you press Enter, the installation program attempts to load each SCSI driver to see whether it can detect your SCSI controller(s). A list of found devices appears as they are detected.

Assuming that auto-detection has found your controller or controllers, you now can press Enter to proceed with the rest of the installation, unless you want to install support for additional devices.


I've seen many situations in which an IDE adapter for an IDE CD-ROM drive was not detected by auto-detection. This is something that you can choose from the list of SCSI controllers. Even though the list is presented as a list of SCSI controllers, there is a choice, ATAPI 1.2, that is the proper choice for an IDE interface. It is not necessary to specify IDE for an IDE hard drive because this driver always is loaded during the installation routine, even if you do not have an IDE hard drive.

After you complete the process of identifying devices and you press Enter to continue the installation, additional drivers are loaded. This includes support for the NTFS file system, support for the CDFS file system, and, if appropriate, support for a SCSI CD-ROM.

Deciding Where to Install Windows NT Server

After your boot drive is examined, a search for a previous installation of Microsoft Windows NT Server is conducted. If a previously installed version of Microsoft Windows NT Server is found, you are asked whether you want to upgrade or install a fresh copy in its own directory.

Assuming that you do not have Microsoft Windows NT Server already installed on this server, you are asked to identify your type of computer, video display, mouse, keyboard, and keyboard layout.

These choices are very similar to the choices you have when running the old Windows 3.x setup routine. Because Microsoft Windows NT Server is so hardware-specific, however, it is very important that you do not make a wrong choice for any of these parameters.

Chances are that the appropriate choices for your server will be shown on-screen, although the video display usually defaults to VGA. I recommend that you leave this choice alone, as well as any other choices that you can, because the first and foremost task that you are trying to complete here is to make your server bootable with Microsoft Windows NT Server. You always can change the video display parameters after Microsoft Windows NT Server is installed.

If you are satisfied with the current choices shown, press Enter to proceed with the installation.

If a copy of Microsoft Windows 3.x or Microsoft Windows 95 is found on your boot drive, the installation routine asks you whether Microsoft Windows NT Server should be installed in the same directory.

It is possible to have Microsoft Windows NT Server and Microsoft Windows 3.x coexist in the same directory, sharing applications and settings. However, there is currently no way to have Microsoft Windows NT Server migrate Microsoft Windows 95 registry settings, which might result in the inability to run the applications you had installed in Microsoft Windows 95 on Microsoft Windows NT Server without the need to reinstall the applications again under Microsoft Windows NT Server.

So, it is strongly advised to install Microsoft Windows NT Server into it's own directory if you wish to continue to run Microsoft Windows 95 on the same PC. You will then need to reinstall the software for use by Microsoft Windows NT Server, if you wish to use the applications from there.

If you do decide to install to the same directory, keep in mind that it might be difficult to uninstall Microsoft Windows NT Server and leave your old Windows intact. I therefore recommend, in all cases, installing Microsoft Windows NT Server in its own directory.

If an old copy of Windows is found, you must press the N key to proceed without installing Microsoft Windows NT Server to the same directory. Press Enter if you do want to install Microsoft Windows NT Server to the same directory as Windows.

If you have not yet told the installation program where to install Microsoft Windows NT Server, you now will find a list of available partitions for installing Microsoft Windows NT Server. It is from here that you can choose to delete an existing partition, and you can create partitions from here as well.

The list shows the drive letter assigned to the partition, the current file system for that partition, and the total space and free space on the partition. Unpartitioned areas of your hard drive(s) also are shown. If you have decided to create or delete partitions, complete those tasks now.

Next, position the highlight bar over the partition to which you want to install Microsoft Windows NT Server. Remember that this partition must be large enough to accommodate approximately 100MB of files. Press Enter; you are asked whether you want to format the partition as FAT, format it as NTFS, convert an existing FAT partition to NTFS, or leave an existing FAT intact with no formatting.

Next, the installation program asks for a directory name. By default, the name is \WINNT, but you can change this if you want to.

After you select a directory name and press Enter, the installation program wants to check existing partitions for corruption. You can allow the program to perform an exhaustive secondary examination of those partitions. Press Enter to allow for both examinations, or press Esc to perform only the first test.

A please wait screen appears during the examination of the hard drive(s). After this process is completed, the file copy process begins.


This process begins only if you successfully identified the device to which your CD-ROM drive is attached. If the installation program was unable to load the appropriate driver, the program prompts you for Setup Disk 4. Well, there is no Setup Disk 4, so you must abort the installation by pressing the F3 key and confirming the operation. Next, you must find a drivers disk for your SCSI controller and then start the installation process all over again.

If you also specified SCSI drivers to be installed from a manufacturer's disk, you are prompted for that disk during this sequence.

After the copying process finishes, you are prompted to remove the disk from your floppy drive. You then can press Enter to proceed with the installation. You now have completed the character-based part of the Microsoft Windows NT Server installation program.

Using the Setup Wizard

Now your server boots Microsoft Windows NT Server from the hard drive and the graphical portion of the installation begins. You can file away those three disks until the next time you need to install Microsoft Windows NT Server or until you need to use your Emergency Repair disk.

When your server reboots, you are presented with a menu of installed operating systems. The first item is your new installation of Microsoft Windows NT Server 4.0. If this is a dedicated server with no other operating systems, this is your only choice. If you originally had MS-DOS on the boot drive, MS-DOS is listed as the second option on the menu. When installation has completed, you find yet another item on the menu, but you'll examine that later in this section.

The boot menu usually has a time-out of 30 seconds before a choice is made automatically, but for this installation, you immediately are launched into the next part of the installation.

You again see the blue screen that identifies the operating system, version number, and build number. Again, under that information, you see the number of processors available to Windows NT and your physical memory, but this time the multiprocessor kernel is loaded only if your server has more than one processor.

The graphical portion of the installation program now starts by initializing, and then some more files are copied from the CD-ROM to the hard drive. The Setup Wizard (Figure 4.1) then starts its three-part process.

Figure 4.1

The Windows NT Server Setup Wizard.

The first part the Setup Wizard will guide you through is defining information about the computer, including the computer name, licensing information, and determining which of the optional components of Microsoft Windows NT Server you want to install. The second part lets you define Microsoft Networking, which includes protocols and services. The final part is setting up your local workspace, which is used only when you logon locally to the server.

The first part is "Gathering information about your computer." Click the Next button. After some subdirectories are created within your Windows NT Server directory, you are prompted to enter your name and the name of your company (if any). This is standard procedure for all Microsoft installation programs. Enter these values and click the Next button.

Choosing Your Client Access Licensing Mode

The Licensing Modes dialog appears (Figure 4.2). As explained in Chapter 3, Windows NT Workstation vs. Windows NT Server, you can license clients for Microsoft Windows NT Server in one of two modes. Per Server mode allows you to specify the number of concurrent users that will be accessing your physical server. A Per Seat mode allows you to purchase licenses for individual users to grant them access to all the servers contained within your corporation. After you specify this mode within the installation program, you have the option to change it from within the installed copy of Microsoft Windows NT Server, but you can change the mode only once during the lifetime of the server.

Figure 4.2

Licensing Mode choice.

Choose the appropriate mode and then click the Next button.

Naming Your Server and Determining Its Role

The dialog shown in Figure 4.3 will be displayed. Here, you need to enter the name of the server. The server's name should not be confused with the name you give your domain. If you plan on having more than one server, you should give the server a name that enables you to easily identify the server when viewing a list of servers. The server's name must contain 15 characters or fewer, and it must be a unique name for your network.

Figure 4.3

Naming your server.

Enter the name and then click the Next button.

The Server Type dialog (Figure 4.4) is where you choose the role for this server. If this is your first server for this domain, you should choose Primary Domain Controller. If you already have established a domain and are adding this server as a backup domain controller that will participate in administrative functions, choose that option. Otherwise, if this server will participate in an existing domain, but you do not want it to participate in logon validation among other administrative tasks, or you do not want to implement a Windows NT domain model and just want to this server to act as a peer-to-peer server, choose the Stand-Alone Server option. In all cases, you must know the name of the domain you will create if you choose Primary Domain Controller, the name of the domain you are joining as a backup domain controller, or the name of the workgroup or domain you will be part of as a stand-alone server. After you make your choice, click the Next button.

Figure 4.4

Choosing the server type.


Keep in mind that during the execution of this Setup Wizard, you can choose to go back to previous screens by clicking the Back button.

All PCs running Windows NT (both Workstation and Server) have an Administrator account. You need to assign a password to that account. If you want to do this later and just assign a blank password for now, you can do that by not entering anything. Do remember to assign a password to the account before you allow other users to attach to this server. Make your entries, if applicable, and then click the Next button.

You are given the choice of creating an Emergency Repair disk. Choose Yes, because you might need that disk to avoid having to reinstall Microsoft Windows NT Server from scratch. Important information regarding your partitions and administrative information is stored on this disk. If you do not make a disk now, or if any information changes that you want to back up to a new Emergency Repair disk, you always can run the RDISK.EXE program from a command prompt.

Selecting Optional Components

After clicking the Next button, you have the option to install optional components for Microsoft Windows NT Server. These items are really for use by a local logon, and are not server tools. These optional components include Accessibility Options, Accessories, Communications, Games, Microsoft Exchange (which is not the Microsoft Exchange Server or client, but a trimmed down version of the Microsoft Exchange Server client that can be used to attach to a Microsoft Mail For PC Networks post office, or for mail from online services), and Multimedia. Check or uncheck the appropriate boxes and click the Next button.

That finishes up part one of the Setup Wizard (Figure 4.5). Part two handles setting up the networking components of Microsoft Windows NT Server. Click the Next button to get started.

Figure 4.5

Finishing part one of the Setup Wizard.

Setting Up Microsoft Networking

The first question the Setup Wizard asks is how your server is connected to the network. Hard-Wired, Remote, or Both are the three choices. Choose the appropriate answer. This section assumes that you are selecting the hard-wired scenario because most of the servers are going to be on a physical network. If this were a server that had to dial in to join its domain, then you would be prompted to answer questions about modems and other parameters, but in this step-by-step process, you will deal with a server that is on a hard-wired network.

Make your choice and click the Next button.

Next you are asked if this server will be used as a server on the Internet. If so, check the box that says you want to install the Microsoft Internet Information Server. This is a fully functional World Wide Web (WWW), Gopher, and File Transfer Protocol (FTP) server, that was formerly available as an add-on for Microsoft Windows NT Server 3.51.

Figure 4.6

Choosing A Network Adapter

As shown in Figure 4.6, the Network Adapter Card section is next. Again, an auto-detect feature can attempt to recognize your installed network interface card. As was the SCSI identification sequence, this can be a point of failure for the installation routine.


If you choose to allow for auto-detection and your server freezes up, you must start the installation process from the beginning.

You also can choose to have a list displayed that you can choose from, or you can use a disk provided by the network interface card manufacturer. If you want to use auto-detect, click the Start Search button. Click Select from List if you prefer to choose your card manually or if the auto-detect does not pick up on the card you have installed.

If you are using a disk provided by a hardware manufacturer, choose Select from List. The resulting dialog includes a Have Disk button. Use this option and the Setup Wizard asks you for the location of the disk.

After you add one or more cards to the list, click the Next button to select the protocols you want to install on your server.

NWLink, which is Microsoft's IPX/SPX compatible protocol, and NetBEUI, which was invented by IBM and is a fast non-routable protocol, are chosen as the default protocols for your server. If you do not want to install these, but you want to choose the TCP/IP option, clear and check the appropriate boxes. You can install as many protocols as you want, but the more protocols you do install, the more overhead there is on your server.

The Select from List button offers even more protocols, such as Microsoft's DLC protocol, which might be necessary if you plan on having 3270 connectivity that might require that protocol, or if you want to see a Hewlett-Packard Jet Direct card on your network. A Have Disk button also is available on this dialog because you might want to install third-party implementations of TCP/IP or other protocols.

After resolving your protocols, click the Next button to display a list of available network services that come with the Microsoft Windows NT Server package. The boxes that appear grayed out are that way because those services are necessary to properly run the NT server, but you can choose the option of installing Remote Access Services (RAS) here. A Select from List button again gives you additional choices.


All these network choices can be made after the installation of Microsoft Windows NT Server, so if you want to try to get the server running with minimum overhead, just accept the default settings to get you through the installation with less possibility of something going wrong.

After the network services are resolved, click the Next button to start the file copy and network configuration process. This is the perfect time to click the Back button to review your choices. If you are using a network adapter card that requires switch settings, you soon are prompted for the IRQ, DMA, and memory address for that card. Other configurable options also are presented to you, but be sure that you know the settings for your network adapter card because the network service attempts to start during this installation process.

Any configuration dialogs can appear, so answer them accordingly. For instance, if you chose to install TCP/IP, you will be prompted to enter your IP address and subnet mask. If you are unsure what any of the right answers are, you can choose to proceed and clean up these network configurations after the installation of Microsoft Windows NT Server finishes.

A dialog appears that allows you to select and deselect and bindings that have created for you by default. Once you have completed verifying the bindings, click the Next button.

After the configuration and file copying process has completed, the network services attempt to start. If they start properly, you proceed to the next dialog, but if they do not start properly, click the Back button a few times to go back to the network interface card dialog and check your settings again. Chances are that the network startup will fail if you gave NT the wrong settings for your network adapter card or if you chose the wrong driver for your network adapter card.

Assuming that you have properly identified the card and its settings, you now should be prompted for the name of the domain that you are starting or joining.

Shown in Figure 4.7, for a primary domain controller you must enter the name of the domain you will be starting. There must not be another domain on your network with the same name. If the network services started properly and you do name a preexisting domain, you will get a warning about this, but it is possible to name an existing domain if you have not been able to connect to the network, so be sure that you have predetermined the name of the domain that you will be creating. If you chose to be a backup domain controller, you must supply the name of the domain that you are going to join. You also must provide the Administrator password, which verifies your authority to add this computer to the domain.

Figure 4.7

Specifying a new domain.

As a stand-alone server, it is not necessary for you to name a domain, but if you do want to join a preexisting domain, do it from here. The domain name, unless it is new, will be validated. If the domain cannot be found, you will not be able to proceed.

Clicking the Next button sends the Setup Wizard searching across the network for a preexisting domain or, in the case of setting up a primary domain controller, validates that the domain you are creating does not exist.

This concludes the network setup part of the Setup Wizard, which brings you to the third and final stage of Microsoft Windows NT Server installation (Figure 4.8).

Figure 4.8

The Final Step For Setup Wizard

IIS and Setting Up the Local Workspace

If you had selected to install the Microsoft Internet Information Server (IIS), this is where you will configure it. Choose the services that you wish to run (World Wide Web, Gopher, and FTP), then click on the next button. Here you specify the directories that you want the selected services to reside in. By default, they will be placed within an INETSVR directory within the Windows NT Server directory. The necessary files for the IIS services you selected are now copied to the server, and each service is started after those files are copied. In the case of the Gopher service, you will receive a message that says, "You do not have an Internet domain name declared for this machine. To insure proper gopher functionality, an Internet domain name should be configured through the Network Control Panel Applet." This can be configured after installation has completed and you restart the server. Clicking on the OK button allows the IIS setup to continue. Once the IIS services have been installed, you are prompted to install ODBC drivers. Microsoft Windows NT Server ships with drivers for Microsoft SQL Server, which you can choose to install at this time.

Windows NT Setup continues by creating default program groups, and performs other tasks to help set up the default environment for Microsoft Windows NT Server.

The first item you are prompted for is your time zone. As shown in Figure 4.9, a graphical map of the world is shown, and all you need to do is choose from a list of time zones. If you are preparing this server for another location that has another time zone, choose the one that is applicable.

Figure 4.9

Setting the Time Zone.

Click the Date & Time tab to adjust the calendar and time settings for your server. Then click the Close button.

Next, the Setup Wizard tells you whether auto-detect has recognized your video display chip. If it has, it asks you to acknowledge this. If it has not detected it, then stick with standard VGA, because you do not want to have to boot to an incomprehensible video display after installation has completed. You can click the Change Display Type button to see a list of video chips that are supported on NT, and you always have the option to click the Have Disk button to use a hardware manufacturer supplied video driver. As I've said before, however, you want to get through this installation process without incident. Changing the display type is something that you can do after Microsoft Windows NT Server is installed. In fact, when you reboot, you will see that another menu item has been added to the boot menu that enables you to boot NT Server in standard VGA mode, just in case you are having a video driver problem.

Click the OK button, set the horizontal and vertical resolution by moving the slider, adjust the color depth, or click the List All Modes button to choose from a list of available resolutions and color-depth combinations. You also can set the refresh frequency here, and you can specify your preference for using large or small fonts in window titles and menu items.

Click the Test button to let NT check the setting you have specified. After the test has completed, you can click the OK button to exit the video display configuration dialog.

Next, the Emergency Repair disk is built. You are prompted to insert a 3 1/2-inch disk into your drive A. The disk will be formatted, so just make sure that it is not write protected. After the format, a number of files are copied to the disk. When it is finished, remove the disk, label it properly, and keep it in a location you will remember if you ever need to use it or update it.

If you chose to install the Microsoft Exchange client, those files now are copied to your hard drive.

And that's it. Your server is ready to reboot and run Microsoft Windows NT Server. Don't forget to make sure that the Emergency Repair disk is removed from your drive A. Click the Restart Computer button, and your Microsoft Windows NT Server starts up ready for you to log in as Administrator, using the password you specified during setup.

You then can create users, create groups, share drives or directories, create print queues, and do whatever it is that you wanted to do with Microsoft Windows NT Server in the first place.

If you had specified that a partition would be formatted as NTFS or converted to NTFS, it is during this first reboot that the NTFS is created. During the installation, you were still working with FAT. You'll see the conversion taking place, so keep your eye out for it.

Good luck!

Examining Installation Problems

Sometimes a Microsoft Windows NT Server installation will go as smooth as silk. Then there's the other times.

I've run into a few problems in the past, so I'll share those with you now. From what I've learned, I haven't had a problem yet installing Microsoft Windows NT Server 4.0, but my experiences with Microsoft Windows NT Server 3.1, 3.5, and 3.51 have had some pitfalls.

Microsoft Windows NT as an operating system is terribly critical of your hardware. Although you might not have a problem booting MS-DOS or Windows 95, NT might not be as forgiving.

Once I was installing an NT server and just after Setup Disk 2 was inserted and the operating system identified itself, a list of HEX codes flashed on-screen, followed by an explosion of ASCII characters in all different colors blinking on and off. I repeated the installation, and at the same place the exact same thing happened. But this time I was better prepared and stared at the screen toward the top to see whether I could see any type of error message before the multi-colored smiley faces appeared.

What I noticed was mention of a parity error. Well, this usually indicates a memory problem, so I ran the hardware manufacturer's diagnostics to see whether I could spot some bad memory. No such luck; it all checked out just fine. I thought I'd install DOS so that I could run some of the DOS-based hardware diagnostic software I had. I installed DOS without a hitch. I ran all the major diagnostic software that was available at the time, but nothing was found.

Meanwhile, I had gotten in several identical servers, so I started to run the install on one of the other servers. A coworker was installing Novell NetWare on his exact duplicate of my server as well.

His installation of NetWare went without a hitch. My second server ended up the same way as my first server, exploding with colorful ASCII characters.

The manufacturer was called in and replaced my system board, just to see whether that would take care of it. Both Microsoft and the manufacturer couldn't offer me any other explanation of what had happened. The new system board didn't make a difference.

I had 48MB of memory in these servers, so just for the heck of it, I pulled out the 16MB of memory that had been added by the Value Added Reseller I purchased these servers from. The server itself shipped from the factory with 32MB. And this time, the NT installed perfectly, without any problems whatsoever. Although the additional memory was made by the manufacturer of the server and went through a quality-control process of the highest caliber, this must have been a bad batch.

Oddly enough, but not surprisingly, the NetWare server died within two months due to a memory problem.

The other times I've had problems installing Microsoft Windows NT Server involved my SCSI controller. Under normal circumstances, a popular SCSI controller is identified properly. An improperly terminated SCSI chain, however, can create a condition that results in an Inaccessible Boot Device error during the first reboot after the character portion of the installation program. It is very important to make sure that both ends of the SCSI chain are terminated. Even if DOS or another operating system is able to overlook certain conditions, Windows NT can be affected by even the smallest error.

Another small problem I've seen is during the SCSI-detection routine. Most people don't think of looking on the list of SCSI devices for an ATAPI driver for their IDE-based CD-ROM drive. More often than not, users are faced with a message asking for Disk 4 because their CD-ROM controller was not detected. As more and more high-speed CD-ROM drives are produced with an IDE interface, it is important to remember that if the auto-detection does not see your IDE controller, you should press the S key to bring up the list of SCSI devices and choose the ATAPI 1.2 option to support your CD-ROM drive.

Although the drivers that ship with Microsoft Windows NT Server usually are the latest at the time of shipping, manufacturers do update their drivers from time to time. This is sometimes to reflect changes in the hardware that has been manufactured since the release date of Microsoft Windows NT Server. For this reason, whenever possible, find out whether there are updated drivers for your network interface card, and any other controllers, sound cards, or other devices. There is no guarantee that the drivers included with Microsoft Windows NT Server will work with your particular piece of hardware.

Also, although the Microsoft Windows NT Hardware Compatibility List might list your server, it also is possible that the manufacturer has shipped a model with slightly different hardware that may not be NT compatible. Most of the larger brand names perform testing to ensure that their hardware remains NT compatible, but the smaller players might not be as efficient. Be sure to get a commitment from your dealer or Value Added Reseller regarding NT compatibility. This way, if it turns out that NT cannot install on your server, you have some sort of recourse and you can get your server replaced with one that will give you a smooth installation.

Another problem that can come up is one of the setup floppy disks going bad. This can happen with any floppy disk, so as Murphy's Law would have it, why shouldn't a disk fail during Microsoft Windows NT Server installation?

To create a new set of setup floppy disks, run the WINNT.EXE program from a DOS prompt. This program is located in the \I386 directory of the CD-ROM, but add to that command line the /O switch. This prompts you for three formatted 3 1/2-inch disks, one at a time. When finished, you will have a brand new set of setup disks from which to work.

In order to troubleshoot a difficult installation, I suggest trying to install another operating system, such as DOS, just to be able to perform simple troubleshooting techniques, such as IRQ or base memory conflicts. Although this method will not always turn up problems that NT may see, it can help eliminate some of the most obvious problems and actually might clear the way for more easily identifying NT-specific problems.

You also might want to create a DOS bootable disk for quick troubleshooting.

Performing a RISC Installation

Installing Microsoft Windows NT Server on a RISC-based server is not much different from what I've described in the "Installing Windows NT, Step by Step" section of this chapter. The major difference is how to kick off the installation routine and the partitioning requirements.

Microsoft Windows NT Server is supported on servers using DEC's Alpha chip, a PowerPC chip, or a MIPS chip. These are Advanced RISC Computer (ARC) compliant RISC systems, which are the only type on which Windows NT can run.

The Microsoft Windows NT Server CD-ROM has separate directories for each of these platforms. There is no need for a boot floppy disk, because the installation program runs directly from the CD-ROM.

A minimum 2MB FAT partition is required on these systems for the boot partition; however, the rest of the hard drive space may be configured as NTFS.

If you are upgrading from a previous version of Microsoft Windows NT Server, you can use the WINNT32.EXE program from the directory that identifies the type of RISC chip that is in your server. For a new installation, follow these instructions.

To install Microsoft Windows NT Server, insert the Microsoft Windows NT Server CD-ROM into your CD-ROM drive and restart your server. When the ARC screen appears, choose Run A Program from the menu. From the prompt, type whatever is appropriate to show that you are executing a program from the CD-ROM (cd:, for example) followed by a backslash, the type of system you are running on (MIPS, ALPHA, or PPC), a backslash, and then SETUPLDR.

On a DEC Alpha server, for example, enter the following:


Then press Enter.


Some of the recent DEC AlphaServers include "Install Microsoft Windows NT Server" on their ARC menu.

Instructions appear on-screen. These instructions prompt you to specify where you want to install Microsoft Windows NT Server. If your system partition is large enough (more than 110MB), you can install Microsoft Windows NT Server onto it. Otherwise, only OSLOADER.EXE and HAL.DLL are copied into a subdirectory, \OS\WINNT.

Never delete the files that are in the \OS\WINNT directory. If you do, you can replace them because they also are copied onto your Emergency Repair disk.


Installing Microsoft Windows NT Server requires careful planning. The hardware must be carefully scrutinized to insure that the installation process will run smoothly. If you are installing your first server, the design of your network should include creating naming standards, and should take into consideration the amount of growth that your company may encounter.

Know how you want to set up your file system in advance. Be prepared to provide the installation program the necessary parameters, such as a computer name, domain name, server role, type of SCSI adapter, type of network adapter and settings for the network adapter. Know what protocols and additional services you wish to install, such as Remote Access Services (RAS), Dynamic Host Configuration Protocol (DHCP), and the Microsoft Internet Information Server (IIS).

In Chapter 5, Introduction to the Windows 95 Interface, you will learn how to navigate around Microsoft Windows NT Server's new GUI. If you are familiar with Windows 95, you already know this, but if you have only previously used the Program Manager/File Manager combination, you are in for a treat.

Previous Page Main Page Page Top Next Page

|  About us | Categories | New Releases | Most Popular | Web Tutorial | Free Download | Drivers |

2019 Soft Lookup Corp. Privacy Statement