Three years later, you’re still consulting for CNT Books. The network has more t
ID: 3787627 • Letter: T
Question
Three years later, you’re still consulting for CNT Books. The network has more than 15 subnets and 10 routers in several buildings and locations. You have been keeping up with the network by configuring the routers statically. However, users have had problems with downtime in the past year because of network links going offline, as there’s only one route to reach every subnet. The owner wants fault tolerance built into the network to include backup links in case a primary link goes offline. You’re concerned that the current router configuration method will still cause some downtime, even if the backup links operate correctly. Why might there be downtime if a primary link goes offline but the backup link is okay? What can you do to reduce the possibility of downtime?
Explanation / Answer
If you're in charge for technical support on a network of any size at all, you've probably discovered that you must make the most efficient use of your time to keep from being completely inundated. One way to accomplish this is to become down to business. Murphy's Law dictates that each computer on your network will fail at some point. So, you should make your workstations as easy to fix as potential to minimize downtime when problems occur.
If you can minimize users' downtime, you'll receive less hate mail from them and free expensive time for more central tasks. In this article, we'll discuss some techniques you can use to minimize the length of time a workstation stays down after a serious run into.
The general idea
The easiest way to ensure that you can fast recover from a crash is to simplify the contents of the PC. In a nutshell, each PC's configuration should be as generic as possible. All profiles, logon scripts, user files, and accounts should be stored on the server.
If you organize your net in this mode and a workstation crashes, you can simply reload Windows NT from the server. You won't have toward worry about spending valuable time performing the complicated recovery techniques needed to save a special configuration or a user's files. Even better, because each workstation has a generic pattern, a user whose PC has crashed can easily work on another PC while waiting for his or her computer to be set.
In an ideal situation, you could make simpler things even more by providing all users with identical hardware. By doing so, you could keep a minimal set of drivers on hand and quickly reload a PC without having to doubt what type of video card or network card was in that particular machine. Unfortunately, as nice as this approach may noise, it doesn't work in the long run. If you upgrade one PC, you must upgrade them all to maintain identical hardware. From a cost and labor standpoint, doing so is impractical. You might think you could return every PC every three years to make sure everyone kept identical hardware. But, a year after you replaced all the PCs, your company might hire new people and PCs identical to those currently in use might not be available for purchase. For these reasons, don't worry about identical hardware. Instead, contemplate on factors that you can control.
Moving files to the server
The first step in making everyone's pattern more generic is to move all user files off the workstation and onto the server. While it may not sound wise to put all your spawn in one basket, keep in mind that your server is probably backed up nightly. Workstations are hardly ever if ever backed up.Start with a very large volume that you'll use only for user files. Take care to construct this volume in such a way that you'll be able to add to its size if necessary. For example, you can add extra hard drives to many hardware-implemented RAID devices to increase the size of the partition with no having to rebuild it. Be sure to format this partition as NTFS.
Granting access to files
Once you have a sufficiently large NTFS divider, you'll need to create a directory that each user can use to store his or her files. We advocate creating a directory called Users and placing subdirectories beneath it, as shown in Figure A. You don't want to bog downhill your server (or your administrative staff) with the task of organization a large number of shares. Instead, create one share at the Users level and grant each one full control of it. You can then create subdirectories that equal each user's logon name and assign these subdirectories NTFS permissions.
User profiles
After you've stimulated everyone's files to a server, it's time to configure profiles. As you may already know, when someone logs on to a Windows NT workstation in favor of the first time, NT creates a profile for that user. NT creates this profile in the WinrootProfiles directory, someplace Winroot is the location of the NT system files. Beneath the Profiles index is a separate directory for each user
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.