Think your server or desktop computer is running up to par? Maybe
it is, and maybe it’s not. But in any case, it pays to check. Neal
Nelson, president of Neal Nelson & Associates, tells you how to go
about it. In some ways a computer is like a car because a small amount of
regular maintenance will dramatically improve its performance and
extend its useful life.
This article outlines six basic steps to a check the health of a
Windows computer. It is necessary to have "Administrator" permissions
before these diagnostic steps can be performed. Microsoft has revised
the menu tree and moved programs around on the various versions of
Windows. I will be giving the menu sequence for Windows 2000. It may be
necessary to use another sequence on the particular version of Windows
that you are running.
I suggest that you bring the machine up to a "normal" state before
you perform the following steps. If you normally have copies of Word,
Excel, Outlook, PowerPoint, Acrobat, Media Player, Quick Books and six
Web browser windows all running and minimized at the same time, then
re-create this load before starting the check-up.
Step 1: Observe the Disk Activity Light
There is almost always a disk activity indicator light on a computer
system. This light goes on whenever there are reads or writes taking
place to the computer’s disk. Many times there is a picture next to the
light that looks like a small beer can. A great deal can be learned by
observing this light. If, during the computer’s boot sequence, the
light is constantly on the computer’s memory might be too small or the
data on the computer’s disk drive might be severely fragmented. If the
light is constantly on, or furiously flashing, during normal operation
the computer’s memory might be too small (causing swapping) or the
memory size is adequate and the machine is a very busy Web, file,
e-mail or database server. For "normal" operation of a workstation
computer, the light should be off most of the time with occasional
flashes as programs access the disk. It is easy to glance at the light
from time to time and use this as an early warning sign that something
is wrong and more investigation is required.
Step 2: Check the Event Viewer
Microsoft maintains a log of events that it feels are either
troublesome or noteworthy. Microsoft also provides a program to display
the entries in this log. It is called the event viewer. On Windows 2000
the menu sequence to run this program is
"Start->Settings->Control Panel->Administrative
Tools->Event Viewer." This will open a window with a scrollable list
of events. The most recent events are at the top. Double click on any
event to open a small "Event Properties" window that displays a more
detailed description of the event or error. The Event Properties window
includes up and down arrows that allow you to scroll through the events
and view the detailed descriptions. Pay particular attention to errors
(marked with a red X) and watch for words like "Harddisk … has a bad
block". Consult with your hardware repair resource about any messages
of this type.
Read more here about how to check the health of a Unix/Linux server.
Step 3: Check Task Manager – Memory Size
Microsoft also supplies a program called Task Manager
that provides a quick display of the system status. Run this program by
clicking the right mouse button while the mouse pointer is positioned
on the task bar (normally at the bottom of the screen) then select
"Task Manager". Click on the "Performance" tab which will display two
sets of graphs and some tables of numbers. Make a notation of the
number listed for "Physical Memory (K) – Total". This is the total
physical memory (RAM) in the computer. Next note the amount of memory
currently being used, "Commit Charge (K) – Total", and the maximum
memory used at any point since the last boot, "Commit Charge (K) -
Peak". If the Commit Charge - Total (memory in use) is greater than the
Physical Memory – Total, the machine is currently swapping or paging.
If the Commit Charge – Peak is greater than the Physical Memory –
Total, the machine went into a swapping state at least once since the
last boot. Swapping or paging is very bad for computer performance. If
there is swapping, you should seriously consider increasing the
machine’s memory size or decreasing the computer’s task load to
eliminate the swapping. Swapping consumes significant CPU resources so
if swapping is taking place don’t bother to perform any other analyses
until the swapping is corrected.
Step 4: Check Task Manager – Percent CPU Busy
The top two graphs on the Performance Tab of the Task Manager show
the percent busy for the computer’s central processing unit. A healthy
and idle workstation computer will tend toward 0 percent busy with
occasional spikes of activity. A healthy server may have a constant
fairly low percent busy with more frequent spikes. A continuous high
percent busy requires investigation. The first step to investigate a
high percent busy condition is to click on the "Applications" tab of
the Task Manager. Review the names of the activities listed. Look for
any programs that you do not think should be running. If you find any,
consult with your system administrator and possibly have them turned
off. The second step to investigate a high percent busy condition is to
click on the "Processes" tab of the Task Manager window. This is a list
of every low level process that is currently running on the computer.
The names can be quite cryptic. Look at the values in the "CPU Time"
column. These numbers indicate the total CPU time that the process has
used since the program started (or since the last boot in some cases).
Any process consuming a lot of CPU time, (ignore the "System Idle
Process"), is a candidate for investigation. It is often helpful to do
an Internet search for the cryptic process name. You may find a Web
page that explains what that process does and whether or not it can be
safely disabled.
Step 5: Check For Disk File Fragmentation
The information on the computer’s disk drive is stored in files that
are made up of "pages." The pages for any file may be stored in a
contiguous sequential group or they may be divided into multiple chunks
which are called fragments. Access to a severely fragmented file is
very slow when compared to access of a contiguous file. There are
several programs available that will analyze and possibly "defragment"
the files on a disk drive. The first is the standard file defragment
program supplied by Microsoft. This program can be accessed by clicking
Start->Programs->Accessories-> System Tools->Disk
Defragmenter. There are other disk defragmentation programs. These can
be found be searching the Internet. I am fond of a free program called
"jkdefrag." Two other popular defrag programs that have license fees
are diskeeper and O & O Defrag. Note that the complete
defragmentation process could take many hours and you should not
interrupt the process once it has started.
Step 6: Check/Clean the Cooling Fans and Heatsinks
You will almost certainly want to have a qualified technician
help with this step. Many systems have "sleeve bearing" cooling fans
because sleeve bearing fans are cheaper than ball bearing fans. After a
while, sleeve bearings dry out, which causes the fans to slow down or
stop. Also, the fans are normally blowing over a heat sink with tiny
fins. These fins can collect substantial amounts of dust and dirt. In
either case the computer will not be receiving proper cooling, which
will cause it to slow down or stop. With the help of a qualified
technician make sure the fans and heat sinks are clean and functioning
properly.A Note About Disk Cache. Windows uses a part of the main
memory for a disk buffer cache. This cache holds copies of frequently
accessed disk blocks. When these disk blocks are accessed from memory
cache rather than from the physical disk the computer’s overall speed
improves. Windows automatically adjusts the size of this cache based on
the amount of "free" memory. This means that increasing the size of
main memory (RAM) would likely increase the size of disk cache and
could result in improved overall performance.
A few minutes for this type of preventive maintenance can greatly
improve your system’s quality of service and also extend its useful
life.
{mosimage} Neal Nelson has more than 35 years experience
with all aspects of complex computer systems. As the chief developer,
owner and president of an independent hardware and software performance
evaluation firm he has tested more than 500 computer systems.
A Web site with some of his test results can be found at www.worlds-fastest.com. For further information send an e-mail to neal@nna.com.
Knowledge Center 
/ 12 
