Analyzing Virtual Memory

We will continue the tour of the Memory Subsystem by analyzing Virtual Memory.

Your Task

Locate and utilize performance analysis tools to determine characteristics of Virtual Memory and how it is utilized and managed by your version of Linux.

Start your Linux Partition from a fresh boot and open the Terminal. Do not open any other applications at this time. Determine values for each of the following. Provide screenshots of command output.

Configure your top command so that the following fields are displayed in the output:

PID

USER

VIRT

RES

SHR

%MEM

SWAP

CODE

DATA COMMAND

Sort the output by %MEM

1) Explain the output from the entire line that starts with Mem: at the top of the Screen.

2) Explain the output from the entire line that starts with Swap: at the top of the Screen.

3) For the topmost process (If sorted correctly, this will be the one using the most memory) give the data values and the meaning behind the value in each column.

4) For the process from #3 above, describe the relationship between the values reported in the VIRT RES and SWAP columns. Does this relationship indicate LINUX may be using demand paging? Why or why not?

Using the vmstat command, run the command such that the virtual memory statistics are captured once per second with a total of 10 samples.

1) What was the command used?

2) What is average value of the swpd field and what does that mean?

3) What is the size of the free list and what does that mean?

4) Out of the 10 seconds sampled, how many of them had swapping activity? How can you tell and what was the activity?

5) Do the values in your findings above indicate any kind of performance problem such as thrashing or the need for more physical memory? Explain.

Now open some applications and leave them open. i.e. Firefox, multiple office apps, and as many other applications as you can while still maintaining a Linux system that functions(the slower the system the better).

Now repeat all of the data gathering steps from above and record new values.

Document before and after values and create a summary documenting the changes you noticed (i.e. The effect of running applications on performance as verified through your performance analysis). Make a recommendation to add resources or change system parameters as necessary to increase performance under heavy loads. Be sure to use your data values to back up your recommendation(s) if any.