Purpose:  to provide you with hands-on experience in using UNIX performance monitoring utilities such as iostat and vmstat.

Preparation:  study the man pages for the UNIX commands vmstat and iostat. If your experience with UNIX is limited, you should also read over a Unix Guide and the man page on the ps command.

Experiment:

1. Use users and ps -ef to determine if other users and processes are using the computer. You should try to use a computer that is otherwise idle for the experiment.

2. Start a browser.

3. After having started the browser, start iostat and vmstat with reporting periods of 5 seconds and redirect their output to a file.
                                       vmstat   5 >   vmstat_output   &
                             iostat   5 >   iostat_output   &
4. Follow several links from the browser during about 2 minutes.
5. Terminate the vmstat and iostat processes. (Use the kill command to kill a process that is running in the background.)

6. Import the files into a spreadsheet to help you compute the following:
• average CPU utilization during the experiment
• average disk service time for each disk
• average number of IOs/sec for each disk
• average disk utilization per disk
• total CPU busy time
• total disk busy time per disk
• average kilobytes paged in
• average kilobytes paged out

Hand in on paper:

1. A brief description of the computer you ran the experiment on. You should include its IP name (e.g. csci.uark.edu), number of cpu's in the computer, the number and names of the disks as deterimined from the vmstat and iostat commands, and the amount of main memory.
2. The printed output from your vmstat and iostat experiment.
3. The summary of your average calculations from above.

Guidelines:

• For this assignment you may share information about the computer platforms, the use of the vmstat and iostat commands, and the use of a spreadsheet for calculating averages. However, each of you should run your own experiments, calculate your own averages, and turn in your own writeup on paper.