For answers to more of your NT questions, visit our online discussion forums at

Norton Speed Disk for Windows NT
Many users seem to think that NTFS doesn't fragment. In truth, NTFS fragments worse than FAT does, especially when disk occupancy approaches about 70 percent. Executive Software and the Norton group at Symantec have documented such fragmentation.

Several years ago, I visited the Symantec group to discuss NTFS and pagefile defragmentation. The engineers were debating the possibility of accomplishing all this defragmentation in one pass. Norton Speed Disk for Windows NT is the result of this debate.

Speed Disk is a powerful disk defragmenter and optimizer. You can move specific files to the beginning of a disk to improve file-access speed. Also, you can control CPU utilization. Screen 1, page 180, shows the file structure of a fragmented disk that Speed Disk hasn't optimized. Screen 2, page 180, shows the file structure of a defragmented and optimized disk.

As you might expect, Speed Disk takes a while to finish its first run. However, after this initial run, the software's runtimes are reasonable. Check out Speed Disk at

I installed an Adaptec SlimSCSI APA-1460 card on my notebook. Are SCSI CD-ROM drives still available?

Considering the availability of SCSI PC Cards and CardBus cards, I would assume that SCSI CD-ROM drives should also be widely available. However, if you carefully search the Internet, you'll find only a few SCSI CD-ROM drives available. I'm willing to bet that CD-ROM drives will soon use Universal Serial Bus (USB). Keep your card, though. Who knows what might develop?

I just added networking to a Windows NT 4.0 Service Pack 4 (SP4) workstation. When I rebooted, a message stated that the Server service couldn't run because of a lack of sufficient storage space. What went wrong?

To replicate your scenario, I installed networking on my notebook. I've concluded that you have incompatible files. For example, my Adaptec SlimSCSI card installation didn't run smoothly. When I put the notebook into Suspend mode, I found that the SlimSCSI driver didn't measure up to Advanced Power Management (APM) 2.0 specifications and could cause file corruption. You can easily solve your problem by reinstalling SP4.

I'm hearing a lot about Fibre Channel Arbitrated Loop (FC-AL). Why is fibre spelled as it is, and what are the concepts behind FC-AL?

ANSI originally designed the fibre-channel architecture to run on fiber-optic cable. When ANSI changed the standard to support copper cabling, an International Organization for Standardization (ISO) task force renamed the technology fibre channel to reduce the technology's association with fiber-optic cable.

Fibre channel offers a SCSI-based command set configured with a gigabit-per-second data-transfer interface that is related to the IP transport protocol. This configuration combines fast I/O with network functionality. Unlike conventional network protocols, fibre channel works in loops that can interconnect at distances as great as 10km.

ANSI gave fibre channel an ISO Level 1 media access control (MAC) layer protocol over which any ISO Level 2 layer protocol can run. For most networking stacks, a datagram or data packet simply maps directly onto the fibre-channel sequence. Asynchronous transfer mode (ATM), IP, and SCSI all map onto the fibre-channel layer protocol.

Early in fibre channel's development, ANSI established several topologies. The cheapest topology to implement was the Arbitrated Loop. This loop connects to a computer's backplane in two places, letting data transfer quickly from node to node (e.g., from computer to computer). Communication relies solely on nodes within the loop. Obviously, such a design has a serious limitation: If one node fails, the entire FC-AL can go down. To avoid this event, you can use Port Bypass Circuits (PBCs) to bypass the node or electronically remove the node from the loop.

One advantage of FC-AL is that it can support as many as 126 nodes on a loop. You can build redundancy into an arbitrated loop, and you can connect independent loops. These capabilities give the loop failover capacity. Another advantage is speed. A single arbitrated loop can transfer data at 100MBps, and a dual arbitrated loop can transfer data at 200MBps. However, real-world environments seldom achieve speeds such as these peak burst rates.

FC-AL is an up-and-coming protocol in the Storage Area Network (SAN) environment. The architecture will probably soon make its way into video streaming and comparable environments.

For more information about FC-AL, see Dean Porter, "Fibre Channel, SCSI, and You," September 1997.

Can you provide a detailed script that automatically runs a system backup?

I assume you're using Windows NT Backup. The basic procedures behind backing up locally and backing up across the network can be somewhat confusing.

The following line backs up the C drive:

                              NTBackup Backup C: /D <"backup description or comment">                                 /B /HC:ON /T normal /L <"%windir%\logfiles\backup.log">                                 /tape0                              

The first portion of this line calls NT Backup and tells it to back up drive C. Next, after the /D parameter, you can write a comment about the backup (e.g., "Backup on November 1"). The /B switch schedules a Registry backup, the /HC:ON switch enables hardware compression, the /T switch dictates a normal backup type (i.e., the system backs up all files), and the /L switch and the path that follows shows that the backup log will write to the logfiles subdirectory of the NT directory. The /tape0 switch is present because the system has more than one tape device. The tape0 switch chooses tape0 (the first device the system records in the Registry) for the backup.

The following line schedules a backup of a shared drive on a remote NT system. (Remember, you can't use NT Backup to back up a remote system's Registry.)

                              net use Z: \\<workstation                                 name>\<sharename><password>                                 /user:<domain name>\backupuser

Because of security permissions, I'm using a user named backupuser. Notice that you need a password in this line.

The next two lines start and stop the backup:

                              NTbackup Backup Z: /A /D                                 <comments> /HC:ON /T normal /L                                 <%windir%\logfiles\backup.log>                                 /tape0

The /A switch tells the system to append data to an existing backup volume, if present. The following line unmounts the shared Z drive (i.e., removes the share from the system):

                              net use Z: /delete                              

Combine the lines you need from the previous examples into a batch file named, for instance, backup.bat. You can then use the At command to schedule your backup.

                              at \\<machine name> 2:00 /Every:Friday backup.bat                              

If I rename the domain name of a BDC in domain A to that of domain B, will the renamed BDC from domain A join domain B?

A BDC in a domain has the same SID as the domain's PDC, so your scenario would fail for the same reason that it would fail on the PDC: The SID associated with the domain controller's original domain is different from the SID of the domain you're attempting to join, even though the domain name is the same. Therefore, you can join an existing domain controller to a different domain only by reinstalling NT on that server.

A utility that might help you is FastLane Technologies' DC Mover, which is part of DM/Manager. Although the utility can't bypass the requirement of reinstalling a domain controller, it eases the reinstallation process. DC Mover lets you save crucial information about the original domain controller and later remigrate that information to the same machine after you've reinstalled NT and joined the new domain. For more information, see Sean Daily, "How to Rename Your NT Domain," February 1999.

Can I remove the arrows from shortcut icons? Also, can I remove the Documents and Favorites submenus from the Start menu?

You can modify the appearance of shortcut icons with the TweakUI utility, which is part of Microsoft PowerToys (available for free download from Microsoft's Web site). To modify the appearance, configure the shortcut overlay setting on TweakUI's Explorer tab. You can choose from an arrow (default), a light arrow, no arrow, or a custom shortcut overlay. For more information about the TweakUI utility, see Mark Minasi, This Old Resource Kit, October 1998.

You can't remove the Documents submenu from the Start menu, but you can alter it so that users don't see the submenu. The Documents submenu is a reflection of the logged-on user's Recent folder (contained in C:\winnt\profiles\username). To restrict the user from viewing the Recent folder's contents, you can delete the contents, then set the NTFS permissions on the Recent folder to, for example, System - Full Control or Creator Owner - Full Control. This fix effectively hides the Documents submenu.

You can directly control the appearance of the Favorites submenu (which exists if you've installed Microsoft Internet Explorer—IE—4.0 or later and Active Desktop) via a Registry entry. Go to the HKEY_CURRENT_USER\Software\Microsoft\ Windows\CurrentVersion\Policies\Explorer Registry key and edit (or add, if it doesn't already exist) the value NoFavoritesMenu. Set the data type to REG_DWORD, and change the value's data from the default 0 (disabled) to 1 (enabled).

Editor's Note: Sean Daily contributed answers to this Tricks & Traps.