InstallationConcepts |
This topic describes concepts and procedures for installing and upgrading to WS2003, installing optional Windows components , and installing third-party applications. Windows Product Activation and Windows Program Compatibility Mode are also covered.
The way you deploy WS2003 depends on several factors:
It's one thing to upgrade two or three servers from NT or W2K to WS2003 in a small company; it's another thing entirely when you have to upgrade thousands of servers across multiple locations in a large enterprise. In the first scenario, you would probably run Setup directly from the product CD, but when the number of servers exceeds about a dozen , automated installations become a more practical solution.
With large numbers of servers having identical hardware configurations, disk imaging is a simple and efficient way of installing or upgrading them. If servers are from a multitude of different vendors and have customized hardware configurations, disk imaging is probably not much of a time-saver.
If only a few staff members are performing the deployment, you need to consider some form of automated installation either using answer files and UDB files or using disk imaging. If the server-to-staff ratio is small, however, it may not be cost-effective to spend the time learning how to perform these types of installations. It may simply be better to install or upgrade from the CD or from a network distribution server.
The standard deployment methods for installing or upgrading to WS2003 include the following:
Here, Setup.exe is run directly from the WS2003 CD to install or upgrade the system. You typically use this method when you have only a few servers to deploy in your network. You are limited mainly by the number of CDs you have and the number of staff members who are available to respond to the prompts as Setup is run.
Setup can also be run over the network by copying the WS2003 installation files to a folder on a file server and sharing the folder. File servers that share the WS2003 installation files are called distribution servers, and a shared folder containing the installation files is called a distribution point. The machines to be installed or upgraded then connect to the distribution point to start Setup, either by creating a network boot disk (see http://www.bootdisk.com) or by mapping a drive to the share. This method can be used to simultaneously deploy dozens or even hundreds of servers. You are limited mainly by the speed of the distribution servers, the network bandwidth available, and the number of staff members available to respond to the prompts as Setup is run.
|
Using a tool called Setup Manager ( Setupmgr.exe ) that's included in \SUPPORT\TOOLS\DEPLOY.CAB on your product CD, you can create answer files and UDB files for performing unattended installations either over the network or from the product CD. An answer file is a specially formatted text file that answers some or all of the prompts during installation. This allows installations and upgrades to be performed without any user interaction other than starting the Setup process. A uniqueness database file, or UDB file, is a specially formatted text file that supplements or overrides some of the information in the answer file. While the answer file provides responses to general prompts, such as which optional components to install or which domain to join, UDB files are typically used to provide system-specific information, such as the names of computers or their IP addresses if DHCP is not being used. You would typically have one answer file for a group of servers in the same department or with the same function, and each installation or upgrade of a server would require its own UDB file. Setup Manager can also be used to create a distribution point for unattended network installs and to create Sysprep.inf files for unattended installs using disk imaging.
Disk imaging (also called disk duplication or disk cloning) is the process of making an exact bit image of a hard drive. You first create a master image of the system/boot disk of a template WS2003 system and then copy or clone this image to other systems. This can be a very efficient method for deploying a large number of new installations of WS2003, but only when the systems you are deploying have identical or very similar hardware configurations. One of the great advantages of disk imaging is that you can use it to install not just a bare-bones version of WS2003 but also a fully loaded server with numerous preinstalled applications. WS2003 includes a utility called the System Preparation Tool ( Sysprep.exe ), which can be used to prepare a system for cloning by ensuring that the cloned systems will have their own unique SIDs. Sysprep works by deleting the SIDS on your existing system, which generates unique SIDs on the target systems when they are restarted after the image has been cloned to them. WS2003 doesn't itself include disk-imaging software, so Sysprep must be used in conjunection with third-party disk-cloning software.
The Windows Preinstallation Environment (WinPE) is a tool available to original equipment manufacturers (OEMs) that can be used to create bootable ISO images of customized WS2003 configurations for rapid, mass deployment of identical systems. If you are a large enterprise customer, you may be able to obtain WinPE from Microsoft and use it instead of disk imaging for mass deployments.
You need to consider some things and make some decisions prior to installing or upgrading a system to WS2003:
Make sure you read Relnotes.htm , Readme1st.txt , and any other last-minute documentation on the product CD concerning installation issues.
You must ensure that all your hardware is fully supported by WS2003. The quickest way to do this is to use the Check System Compatibility feature of Setup, which can be started from the product CD menu. You can also consult the WS2003 Hardware Compatibility List (HCL), which lists devices whose drivers have been tested for and comply with WS2003. Microsoft supports only hardware that is listed on the HCL, so be sure to comply with this list if you want to be eligible for Microsoft's technical support. The HCL can be found on Microsoft's web site at http://www.microsoft.com/hwtest/hcl/.
You must also meet the minimum hardware requirements for installing WS2003, though experience says you should go well beyond the recommended hardware requirements if you want satisfactory performance. Table 4-22 lists the minimum hardware requirements for different editions of WS2003, while Table 4-23 shows recommended hardware. Table 4-24 displays the maximum RAM and supported number of processors for each edition.
Requirement | Standard Edition | Enterprise Edition | Web Edition | Datacenter Edition |
---|---|---|---|---|
CPU speed | 133 MHz | 133 MHz | 133 MHz | 400 MHz |
Memory | 128 MB | 128 MB | 128 MB | 512 MB |
Disk space | 1.5 GB | 1.5 GB | 1.5 GB | 1.5 GB |
Requirement | Standard Edition | Enterprise Edition | Web Edition | Datacenter Edition |
---|---|---|---|---|
CPU speed | 550 MHz | 733 MHz | 550 MHz | 733 MHz |
Memory | 256 MB | 256 MB | 256 MB | 1 GB |
Disk space | 1.5 GB | 1.5 GB | 1.5 GB | 1.5 GB |
Requirement | Standard Edition | Enterprise Edition | Web Edition | Datacenter Edition |
---|---|---|---|---|
Max RAM | 4 GB | 32 GB | 2 GB | 64 -512 GB |
Min CPUs | 1 | 1 | 1 | 8 |
Max CPUs | 4 | 8 | 2 | 32 |
Although a 2.5-GB partition will suffice for installing WS2003, use at least a 4-GB partition to leave room for additional WS2003 components you may want to install later. You can create and delete partitions during the text-based initial portion of Setup, but Microsoft recommends that you use Setup only to create the partition on which you plan to install WS2003 and then use the Disk Management console after the installation is complete to create and format other partitions. Besides speeding up Setup, using Disk Management gives you the option of converting your disk subsystem to dynamic storage so you can create extended and fault-tolerant volumes .
During the text-based portion of Setup, you can specify which filesystem formats the partition on which you install WS2003. You can choose between FAT, FAT32, and NTFS, but the logical choice is NTFS. NTFS provides additional security and manageability through NTFS file and folder permissions, EFS encryption, disk compression, and disk quotas. The only reason you might want to choose FAT or FAT32 is if you want to be able to dual-boot your system, but this is a highly unlikely choice with a production server.
Decisions about how you want to license your server must be made prior to installation. The three aspects of WS2003 Licensing are:
One server license is required to license your right to install and run WS2003 on the computer.
Multiple client-access licenses (CALs) are required to license client computers with the right to connect to your server and access its services. (In addition to determining the number of CALs you require, you also need to decide whether you will license these CALs in a per-server or per-seat mode.)
Additional licenses may be required if you have other Microsoft applications installed and running on your server.
You need to decide whether your server will be installed as a standalone server that is part of a workgroup or as a member server belonging to a domain. If you plan to join the computer to a domain during Setup, then you will need:
The name of the domain you plan to join (e.g., mtit.com ).
A computer account to be created for your computer in the domain you plan to join. This can be done by creating the computer account ahead of time using Active Directory Users and Computers, or you can create the computer account during Setup, provided you have the credentials of an administrator in the domain (member of the Domain Admins group for the domain).
An available domain controller and DNS server for the domain.
If your system is connected to the Internet during Setup, you can use the new Dynamic Update feature of WS2003 to automatically download the latest Setup files and device drivers during installation.
If any existing partitions from a previous operating system have been compressed with DriveSpace or DoubleSpace, make sure you uncompress them. If you have a mirror set from a previous operating system, break the mirror prior to upgrading, then recreate the mirror set after the upgrade is complete. Finally, if you are using a UPS, disconnect it before installing WS2003 on a machine.
After installation is complete, you can log on to your standalone server (if the server belongs to a workgroup) or member server (if the server belongs to a domain) and perform various postinstallation tasks .
You must activate your software after installing it. Windows Product Activation (WPA) is a process that links your product key with your hardware configuration and is used by Microsoft to discourage software piracy. Activation can be performed over the Internet or using the telephone, and no personal information is collectedin fact, the hardware information itself is hashed so that Microsoft doesn't even know what hardware you are using. If you don't activate Windows in the required grace period (30 days), you will be unable to log on to your system, so make sure you use WPA the first time you log on to your new system. Note that customers with enterprise volume licensing agreements don't have to activate their systems.
You should check to see if everything went well during Setup:
Check Event Viewer to see if there are any error or warning messages associated with the installation process. Also, configure your event log settings as desired.
Use Services in Computer Management to check that all services set to Automatic have started successfully.
Use Device Manager in Computer Management to ensure that your hardware devices have been detected properly and assigned appropriate resource settings.
Verify your IP address, DNS, and WINS settings by typing ipconfig /all at the command prompt. Also, try connecting to a shared folder on the network to see if your network connection is working properly.
Additional configuration steps you may want to perform immediately include:
Use Display in the Control Panel to configure your display for at least 800 x 600 screen resolution.
Use Power Options in the Control Panel to configure your ACPI power scheme. Servers should generally have their hard drives configured not to power down during idle times.
Use System in the Control Panel to configure your Startup and Recovery settings as desired.
When you first log on to your newly installed or upgraded system, the Manage Your Server Wizard runs, prompting you to add roles to your server. Supported roles include domain controller, file server, print server, application server, and so on.
Upgrading W2K servers to WS2003 is basically trivial, so this section focuses mainly on upgrading NT servers.
Installing means putting WS2003 on a newly formatted partition or putting it on a partition having another operating system to create a multiboot machine. Upgrading means replacing an earlier operating system with WS2003. Installing means that you have to specify all of the user- and computer-specific settings for your machine, either by answering prompts during Setup or by using answer files to perform automated installations. Upgrading means that the existing user- and computer-specific settings from the previous operating system are carried over as much as possible to WS2003. Upgrading also means that your existing applications don't need to be reinstalled and reconfigured, provided, of course, that these applications are fully compatible with WS2003.
The supported upgrade paths for Standard and Enterprise Editions of WS2003 are shown in Table 4-25. Note also that:
You can upgrade Standard Edition to Enterprise Edition.
You can't upgrade any platform to Web Edition.
If you want to upgrade to Datacenter Edition, you should contact your OEM.
Current platform | Upgrade to Standard Edition | Upgrade to Enterprise Edition |
---|---|---|
NT Server 3.51 | Upgrade to NT 4 first | Upgrade to NT 4 first |
NT Server 4.0 | Yes | Yes |
NT Server 4, Terminal Edition | Yes | Yes |
NT Server 4, Enterprise Edition | No | Yes |
W2K Server | Yes | Yes |
W2K Advanced Server | No | Yes |
When upgrading an NT Server-based network to WS2003, you have a choice of which servers to upgrade first:
This approach provides your network with many of the advantages of WS2003, including new management tools, group policies, support for USB and Plug and Play hardware, an updated version of NTFS with support for encryption and disk quotas, new services, better printing support, and so on. Without Active Directory, however, you must maintain your old NT Server domains until you upgrade your domain controllers. Nevertheless, you may want to choose this method since it gives you many of the advantages of WS2003 while allowing you to buy time to learn the complexity of Active Directory.
This approach immediately provides you with all of the WS2003 features described earlier, plus the power and scalability of having Active Directory on your network. The downside is that you really have to know how Active Directory works before you start implementing it since you don't want to create a directory structure only to have to tear it down and rebuild it later.
The key thing, of course, with upgrading your servers is always to make a full backup of your system before you upgrade. Apart from that, the methods and approaches for upgrading to WS2003 are identical to those for performing clean installs, as discussed previously in this chapter.
After installing WS2003 and adding roles to your server, you can install optional Windows components using Add or Remove Programs in the Control Panel. The additional components you have vary a bit with the OS edition, but generally include:
You can also use Add or Remove Programs to install third-party applications on your server from floppy, CD-ROM, DVD-ROM, or software distribution points on your network. It's best to make sure that your third-party applications are compatible with WS2003 before installing them. Programs that are fully compatible with WS2003 are certified as such through Microsoft's Certified for Windows (CfW) program. See Veritest's web site at cert.veritest.com/CfWreports/server/ for a searchable database of CfW products.
If you're upgrading instead of installing, you probably have legacy applications on your server that were designed for W2K, NT, or even Windows 95/98/Me. In this case, you can use a new feature of WS2003 called Program Compatibility Mode to configure these applications to run as well as possible on the new platform. What you can do is first test your applications to see if they run properly after the OS is upgraded, and if "issues" appear, try running them under one of the following compatibility modes:
You can also configure the display settings to ensure compatibility as follows :
The easiest way to test your programs is to run the Program Compatibility Wizard. Alternatively, you can manually try different compatibility settings to see the result. If your application still doesn't run well after this, you can visit the vendor's web site to see if there is a patch or update for the program to make it work under WS2003. You can also visit the Windows Update web site to ensure your system files and drivers are fully up to date, as driver problems may be affecting your third-party programs.