Information Technology Reference
In-Depth Information
add memory to the server before it crashes; in the past, this process meant shutting down
the server first. Not so with Windows Server 2008 Enterprise and Datacenter editions
because both support
hot-add
memory. Unfortunately, the server hardware must also sup-
port this feature, and you find it only in high-end, enterprise-class servers. Some servers
even support adding or replacing a processor without a system shutdown, but this feature
is supported only in Datacenter Edition. The capability to hot-add disk drives is more
common and can be found in almost all server classes. If you need more disk space or
need to replace a failed disk in a RAID configuration, you can simply install the new disk
without shutting down the server. All editions support disk
hot-replace
or hot-add if the
hardware supports it.
This list covers just a few of the server hardware features you should consider before installing
a new server. The best advice is to forge a good relationship with a knowledgeable vendor you
can consult with when you need to make a purchase. This way, you can focus on managing your
server, and your vendor can focus on keeping up with the latest hardware options.
To make sure hardware selections are compatible with Windows Server 2008,
check the Windows Server Catalog at
www.windowsservercatalog.com.
When installing the first server in a new network, you must make some decisions shortly
after finishing the installation. Some are fairly straightforward, but others take some thought
and consultation. Here's a list of some decisions you need to make:
• What should you name the server? This decision is more important than it sounds. Every
computer needs a name so that it can be identified on the network. A server name must be
unique on the network and should include some description, such as its location or pri-
mary function. Server names should also be simple and easy to remember because users
often access servers by name.
Even if you expect the server to be the only one on the network, you
shouldn't use just “Server” as the name. Situations often change and require
adding a server, so at least give it a number, such as Server1. Subsequent
server names can be a bit more descriptive, such as Mail1, Accounting1, or
Room19.
• Which network protocols and addresses should you use? By default, Windows installs both
TCP/IPv4 and TCP/IPv6 in Windows Server 2008 and Vista. You can't uninstall them, but
you can disable them in a network connection's Properties dialog box. Disabling a protocol
is recommended if you're not using it. TCP/IPv4 is still the predominant LAN protocol and
probably will be for years. Previous Windows versions had the option of installing other
protocols and services, such as IPX/SPX (NWLink) and client/server components for
NetWare. Windows Server 2008 has no additional protocol or client options, so if they are
important, you need to find a third-party solution or use Windows Server 2003 or earlier.
• How should I assign an IP address to the server? By default, Windows Server 2008 uses
automatic IP addressing, but a server should have a static IP address. Some server roles
actually require assigning a static address. If you haven't devised your addressing scheme,
now is the time to do that. Generally, servers use one of the first or last addresses in the
address range, such as 192.168.1.10 or 192.168.1.200. Whatever you decide, be consis-
tent so that when more servers are added, you can assign addresses easily.
• Setting the correct time zone isn't a decision but a task you must complete because having
the wrong time zone can cause all manner of problems, particularly in a domain envi-
ronment. Certain functions in a domain network, such as user authentication, depend on
client and server computers having their clocks well synchronized.
• Should I use the workgroup or domain model? As discussed in Chapter 1, the Windows
domain model has advantages in usability, manageability, and security. If you've invested in
Search WWH ::
Custom Search