Optimization and Troubleshooting

No matter how well your network is running, it can always be improved. Windows 2000 Server includes a lot of monitoring tools. The most common tools include the Task Manager, the Performance Console, and the Network Monitor. After all the hardware is installed and the network is up and running, a good portion of a network administrator’s job is looking for ways to improve performance and addressing potential problems before they get out of hand. You’ll probably become very familiar with all the tools described in this section.

Task Manager

The Windows 2000 Task Manager allows you to monitor the status of individual applications, processes, and both CPU and RAM usage. You can start the Task Manager by right-clicking on the task bar and choosing Task Manager or by pressing Ctrl-Alt-Del while Windows 2000 is running.

Under the Application tab, you can start and stop entire programs. Under the Processes tab, you can view the status of individual processes. The process data shown includes a process ID, the percentage of CPU time being used, and the amount of RAM being used. You can also stop a process from this screen. The performance tab allows you to monitor the CPU and RAM usage for the local computer.

Performance Console

The Performance Console is a container for two snap-ins: the System Monitor snap-in and the Performance Logs and Alerts snap-in. You can monitor both the local computer and remote computers on the network with the System Monitor and the Performance Logs and Alerts snap-in.

The System Monitor can be used for live monitoring, or it can store data to be later presented in graphs, charts, or reports. You should run the System Monitor as soon as your Windows 2000 network is set up properly and users are accessing the network resources without any trouble. This original performance data is called a baseline . Later on, you can compare the current performance to the baseline data to measure changes in the network. This will help take the guesswork out of determining how best to improve and expand your network in the future.

The Performance Logs and Alerts snap-in is a Windows 2000 service that can run and collect data for later analysis in the System Monitor. It can collect data from both local and remote computers. You can run more than one copy of the Performance Logs and Alerts service simultaneously.

Network Monitor

The Windows 2000 Server Network Monitor can monitor, capture, or store data that is being transmitted on the local subnet. The sheer volume of data usually prevents an administrator from wanting to capture all of the data that is passing over the subnet.

You can filter the type of data you’d like Network Monitor to capture. For example, you can monitor all the data passed between two IP addresses or all the data transmitted using a particular protocol. You can also create a set of filtering rules using common logical operators such as AND, OR, and NOT.

Network Monitor can display the data for your analysis or allow you to save it in a log file that can be sent to someone who specializes in analyzing network traffic.

SNMP

Simple Network Management Protocol (SNMP) is a common protocol used to monitor and troubleshoot TCP/IP. SNMP distributes the work among three main components, the Network Management Station (NMS), the agent , and the Management Information Base (MIB):

  • The NMS receives data that was stored in the MIB by the agent.

  • The agent collects data from the monitored hardware and stores it in the MIB.

  • The MIB is a database of stored information collected by the agent meant for analysis after it is transferred to the NMS.

To provide limited security, computers that will be using SNMP can use a community password shared by all the agents and network management stations. This password is called the SNMP community name .

There are several third-party programs that provide a user-friendly way of managing the monitoring of a Windows 2000 network through the use of SNMP. Some of these programs can be set up to email or page system administrators when problems occur.

Disk Tools

Windows 2000 Server includes two disk troubleshooting tools, Check Disk and Disk Defragmenter. Microsoft has almost always included both a check disk and disk defragmenter program with all their operating systems, with one notable exception: Windows NT did not come with its own built-in disk defragmenter.

Windows 2000 includes the Disk Defragmenter snap-in. Windows 2000 Server does not automatically optimize where data is stored on a disk. As files are added, modified, or deleted, the fragmentation of data increases. This results in a performance decline. It is usually quite gradual, so be sure to set up a regular schedule for defragmenting the disks on your network. Be sure you run Disk Defragmenter only when the server isn’t doing much disk I/O to avoid major performance penalties.

Check Disk is a simple tool used to check for both filesystem errors and bad sectors on the hard drive. You should try to close as many applications as possible before running the Check Disk program.

Hardware Issues

Microsoft publishes a Hardware Compatibility List (HCL). There is a copy of this list on the Windows 2000 Server installation media and an updated copy on Microsoft’s web site. The hardware on the HCL has been specifically tested and approved for the Windows 2000 platform. Windows 2000 may run perfectly well on hardware that is not on the HCL, but the HCL gives you an opportunity to let someone else do the testing for you. You may have a more difficult time getting Microsoft to help you with hardware that is not on the HCL.

Safe Mode

Windows 2000 Server can be started with the bare minimum of device drivers and services running to help you troubleshoot a problem. This is particularly useful if you installed a faulty driver or made a severe configuration error. After you’ve fixed the problem, you can restart Windows 2000 normally.

Safe mode allows you to gradually increase the number of drivers and services that are running until you’ve solved the problem. Safe mode can be used with or without networking and in VGA mode. You can also use the Last Known Good Configuration, send debugging information to another computer through a serial cable, or store data locally to a log file during the boot process.

If none of these techniques works, you may have to use the emergency repair disk (ERD) or reinstall Windows 2000. You can create an emergency repair disk by choosing Start Run, typing RDISK , and clicking OK. Be sure to have a blank floppy ready and follow the on-screen instructions.

Recovery Console

The recovery console is a command-line interface used for troubleshooting a damaged Windows 2000 computer. You can run the recovery console by booting the damaged server using either the Windows 2000 installation floppies or, if possible, by booting from the installation CD-ROM. The recovery console allows an administrator to start and stop services as well as providing access to most areas of the local hard drives. You need to log in with the local administrator’s password to use recovery console.

To install the recovery console, you’ll need to run the winnt32 program with the switch /cmdcons, which will then prompt you to authorize the change; from then on, when you boot, another line reading Microsoft Windows 2000 Command Console will be available. You need a local administrator account to use the recovery console.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.21.248.162