Ask Dr. Bob - 11 Sep 2000

I'm attempting to install Microsoft Office 2000 on my company's Windows 2000 and Windows NT 4.0 machines. Although I have administrative rights, the installation process doesn't recognize my privileges and informs me that I must have administrative rights to complete the installation. What is going on?

You've encountered a common Office 2000 problem. While attempting installation, you probably received one of the following messages: Installing this product requires Administrative privileges on systems running this version of Windows or Microsoft Office 2000 Setup needs to replace some Windows components. This requires Administrative rights. Please log in as an administrator to complete Microsoft Office 2000 Setup.

You're probably experiencing this problem because you belong to a large number of user groups and, therefore, the system doesn't recognize your administrative rights. (To view the user groups you belong to, go to Administrative Tools, User Manager for Domains.) The simple solution is to obtain Office 2000 Service Release 1. SR1's installation process fixes the problem. Like the just-released Win2K SR1, Office 2000 SR1 requires you to have the original Office 2000 CD-ROM to perform the upgrade.

I'm attempting to install a Digi International DigiBoard card on my Windows NT 4.0 Service Pack 6 (SP6) system, but the installation doesn't work. BIOS Setup recognizes the card, but NT doesn't. To further complicate matters, the installation of the card usually causes NT to hang at the hardware detection portion of startup. What can I do?

Although the motherboard PCI firmware typically assigns I/O port addresses or other resources to PCI devices, NT's Hardware Abstraction Layer (HAL) automatically assigns an I/O port address or memory resource to the DigiBoard card that overlaps with an existing PCI device. The result is a PCI resource conflict on your system. Unfortunately, the HAL sometimes moves resources even when they shouldn't be moved.

Under ordinary operation, the system BIOS sets the PCI device resource requirements. However, when NT loads, the HAL sometimes moves the PCI device resources. The motherboard can dynamically configure PCI devices.

The problem originates in the device driver. Device drivers typically use HalAssignSlotResources as the API that moves the PCI resources. However, some drivers (including that of the DigiBoard card) use IoReportResourceUsage instead of HalAssignSlotResources and obtain the current resource settings from the PCI device's configuration space. To read the PCI device's configuration space, NT uses the driver's built-in calls to HalGetBusData and HalGetBusDataByOffset. NT uses this call method only as a last resort because it contradicts the purpose of the dynamically configurable bus.

To solve the problem, use the /pcilock switch in boot.ini at the end of the boot line. This switch causes the HAL to stop polling the devices. When the HAL recognizes the /pcilock switch—and the system BIOS settings recognize the device—the HAL doesn't let the OS change the PCI device resources. This solution works only in Microsoft HALs (i.e., not in custom-written HALs). According to Digi International, the /pcilock switch will bring healthy functionality to your DigiBoard card.

If you're experiencing a similar problem on a Windows 2000 machine, the /pcilock switch won't work—probably because Plug and Play (PnP) always polls devices and the switch prevents polling. In this case, you probably need to upgrade to a motherboard that Microsoft's Hardware Compatibility List (HCL) supports.

I'm considering buying a Quantum Snap Server for inexpensive storage. Do you have any experience with some of the new Network Attached Storage (NAS) devices on the market?

I recently bought the Snap Server 4000 for my network, but I never got beyond reading the manual. The problem with many NAS devices is that you can set permissions only at the share level. Therefore, every directory and file under that share has the same permissions as the share. This limitation might be acceptable in some situations but not in networks that value security. However, if these inexpensive NAS servers are sufficient for your purposes, they offer an impressive combination of usability and value.

I purchased a Hewlett-Packard (HP) NetServer LH3 with 512MB of RAM and a 500MHz processor. The system's RAID controller keeps marking my RAID drives—two 9.1GB drives in a RAID 1 configuration and four 18.1GB drives in a RAID 5 configuration—as offline and faulty. I've replaced the drives, but the problem seems to resurface twice a month. Do you know what is going on?

I can think of a few possible causes of your problem. First, check the obvious—bad cable, power problems, or room temperature. (Hard disks, especially some faster disks, can become very hot.) Second, try flashing (i.e., booting to a 3.5" disk that contains the most recent firmware revision) the RAID controller and the Hot-Swap backplane firmware. (You can download the HP firmware from http://www serve6717_swen.html.) The NetServer LH3's Hot-Swap drive cages have their own firmware. Problems can develop with the firmware on the drive backplane, causing drives to go offline intermittently. Third, if flashing the firmware on the RAID controller and backplane doesn't fix the problem, contact HP and get the backplane replaced. Replacing the backplane is a straightforward task, but be sure to keep track of each drive's position in the drive cage. You'll need to replace the drives in their original order. I recommend flashing the backplane firmware with the latest firmware version before you bring the server back online.

I recently added a new backbone switch to my network. I now have five 24-port 10/100 Ethernet switches connected by a gigabit-stacking module. The bottom switch in the stack connects to the new backbone switch through a Gigabit-SX uplink. I moved my email server, file share server, network printers, and a few Sun Microsystems workstations to the backbone switch. Since the move, users are having problems uploading files to the server, and printers are producing blank pages in the middle of print jobs. The Sun workstations are having problems copying files from one system to another, and I'm getting NFS mount errors. I have no problems downloading data from the servers; the errors occur only on upload. Do you know what is causing these errors?

Because your errors occur on more than one operating platform and everything worked fine before you added the new switch, I don't think your problem is OS-related. A few possibilities spring to mind.

First, check the Gigabit-SX transceiver and cable on the bottom switch in the stack and on the new backbone switch. A bad transceiver could be the root of your problem.

If this switch provides an error-logging function, search the logs for errors. Sometimes, 10/100 Ethernet cards have trouble negotiating with the network switch. If your logs show ports going up and down several times a day, try selecting 10MB or 100MB instead of Auto on the network adapter's Properties screen. (However, remember that you'll probably see increased activity in the mornings and evenings, when users turn their PCs on and off.)

If the problem persists, contact your switch vendor or download the latest driver for your network card. I recently encountered a problem in which the 3Com 3C595-TX Ethernet adapter card didn't work with switches from various suppliers. Downloading the most recent 3Com driver corrected the problem.

By the way, you need to be mindful of the number of hops between the originating machine and the destination resource. Some switches have trouble with as few as 6 hops—the number you have in your stack (i.e., five switches and one destination resource). Most switches can handle at least 7 hops, and some can handle 15 or more. To potentially avoid this excessive-hop problem and improve performance, try moving your backbone switch to the center of the stack, which essentially creates two switch stacks. This procedure requires two more gigabit transceivers, one for the backbone switch and one for your second switch stack.

Recently, the Windows NT Server 4.0 machine that I use as a network print server has been giving me problems. The print server hosts Hewlett-Packard (HP) printers, a few QMS 2560 printers, and a Minolta PageWorks 18L printer. The spooler service crashes sporadically, causing an interruption in companywide printing. Any suggestions?

I've encountered this problem before, and I located the cause in the Minolta printer. The PageWorks printer driver doesn't always work well with NT's print spooler. When you send a print request to this printer, the printer launches a print-queue-monitoring application called SuperRIP, which tends to crash NT's spooler service. To work around the problem, try installing the IP Peer-to-Peer printing software that ships with the Minolta printer. Then, you can disable SuperRIP by clicking the Advanced tab on the printer's Properties page and selecting the Disable Print Monitoring check box. The Print Monitoring screen will no longer pop up, but this fix offers the additional benefit of speeding up printing.

You can install the IP Peer-to-Peer software on the server or the client. If you install the software on the client, the client can print through the network directly to a printer. You can also use the IP Peer-to-Peer software on QMS printers.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.