How do I measure NFM-P performance?
Steps
1 |
Log into a main server, auxiliary server, or main database station as the root user. |
2 |
Open a console window. |
3 |
Enter the following to list the processes that have the highest CPU usage: # top ↵ Depending on your system configuration, approximately the top 20 processes are displayed. The top NFM-P process, by CPU usage, is typically the Java process. |
4 |
Review the output; see the top man page for a description of the output fields. |
5 |
Record the data for future performance comparison. Look for data that indicates excessive or continuously increasing CPU usage by the Java process. |
6 |
Press Ctrl+C to stop the command. |
7 |
Enter the following to list CPU resource usage information: # mpstat nn ↵ where nn is the time, in seconds, between CPU polls; a value between 10 and 60 is recommended |
8 |
Review the command output; see the mpstat man page for a description of the output fields. |
9 |
Record the data for future performance comparison. Look for a difference in the output for a similar load on each station; a difference may indicate CPU performance degradation. |
10 |
Press Ctrl+C to stop the command. |
11 |
Enter the following to list disk performance information: # iostat -x n ↵ where nn is the time, in seconds, during which you want to collect data; a starting value of 2 is recommended |
12 |
Review the output; see the iostat man page for a description of the output fields. |
13 |
Record the data for future performance comparison. Look for a difference in the output for a similar load on each station that may indicate disk performance degradation on a station. |
14 |
Press Ctrl+C to stop the command. |
15 |
Use an OS command to list network interface throughput and error statistics. |
16 |
Record the data for future performance comparison. Look for a difference in the output for a similar load on each station that may indicate network performance degradation. |
17 |
Close the console window. End of steps |