Linux System performance monitoring tool Nmon

First, Introduction

The Nmon tool helps to display all of the important performance optimization information on one screen and dynamically updates it. This efficient tool can work on any dumb screen, Telnet session, or even dial-up lines. In addition, it does not consume a lot of CPU cycles, usually less than 2%. On the updated computer, its CPU utilization will be less than 1%. Use a dumb screen to display the data on the screen and update it every two seconds. However, you can easily change this interval to a longer or shorter time period. If you stretch the window and display the data in x Windows, VNC, PuTTY, or similar windows, the Nmon tool can output a large amount of information at the same time. The Nmon tool can also capture the same data to a text file for later analysis and drawing of the report, with the format of the spreadsheet in the output file.

Second, download

: Http://nmon.sourceforge.net/pmwiki.php?n=Site.Download
The downloaded file name is: nmon16d_x86.tar.gz (different download times, the version number may be different)
Download the file to local, unzip, unzip the contents of the folder as shown in the following section:

Third, the use
According to your own Linux system version in the extracted folder after the selection of appropriate files, generally with nmon_linux_x86_64 can. Then use FTP to upload to the Linux server's/usr/local/bin directory and add executable permissions.
Modify permissions: chmod 755 nmon_linux_x86_64
Input nmon_linux_x86_64 in any directory can display CPU and other information,
(Note: If you feel nmon_linux_x86_64 this command is too long, you can change the file in the/usr/local/bin directory to the name MV Nmon_linux_x86_64 Nmon)
Enter Nmon in any directory at this point, as shown below:

Iv. Generating Nmon Reports

#nmon-s10-c30-f-m/home/chen

Parameter explanation:
-S10 collects data every 10 seconds;
-C30 Acquisition 30 times, total time: 30*10=300 seconds;
The report file name generated by-F contains the time the file was created;
The storage directory of the report file generated by-m;
The resulting file name is as follows: Xxxxx_180816_1059.nmon, "XXXXX" is the hostname of the machine.

You can see more ways to use nmon-h with Nmon.

V. Use the Nmon_analyser tool to view the report

Upload generated reports locally and view them with the Nmon_analyser tool
1. FTP The Xxxxx_180816_1059.nmon file under the/home/chen directory to the local folder
2. Download Nmon_analyser Tools
Address: Https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/Power%20Systems/page/nmon_analyser
File name to download: Nmon_analyser_v54.zip (Note: The file name may vary depending on the download time)
After downloading to the local decompression, open the. xlsm file, click ' Analyze nmon Data ', you may be prompted ' macro is not available ', click ' Enable Macros ';
Click on ' Analyze nmon data ' again to select the. nmon file that was just ftp to the local, as shown in:

Vi. Nmon Results File analysis
Nmon file parsed out after the generated Excel table, there are dozens of sheet; The following is an analysis of the meanings of the indicators listed in these sheet.

Analysis of common indicators
Monitoring operating system, the most concerned about the indicators are CPU, memory, disk IO, network and so on. Several key charts are analyzed below.

A.cpu_all: Overview of all CPUs, showing the average occupancy of all CPUs in the monitoring system, including the User/sys/wait/idle status
1, user%, percentage of CPU used by programs executed in user mode
2, sys%, percentage of CPU used by programs executing in kernel mode
3, wait%, the percentage of time spent waiting for IO
4, IDEL%,CPU of idle time percentage, this value and user%,sys%,wait% sum equals 1
5, CPU%,CPU Overall occupancy, this value is usually equal to user%+sys%+wait%
6, cpus,cpu the number of cores, that is, the operating system is how much C

B. Disk_sum: Overall DISK read, write, and I/O operations
1. Disk read kb/s, performing sampled data per disk (read rate of disk device)
2. Disk write kb/s, each disk performs sampled data (disk device write rate)
3, Io/sec, the number of transfers per second output to the physical disk

C. diskbusy: Average occupancy of each hdisk device
Unit is% (percent)

D. MEM: Memory usage description, including physical memory and virtual memory
1, Memtotal, total physical memory size
2, Swaptotal, total size of virtual memory (i.e. swap space)
3, Memfree, the remaining physical memory size
4, Swapfree, the remaining virtual memory size
5, cached, file system cache size occupied, allocated by physical memory
6, buffers, file system buffer size
7. swapcached, the amount of memory allocated in virtual memory
8. Inactive, the memory size that is not used most recently

E. NET: Data transfer rate per network adapter in the system (kilobytes per second)
1, Total-read, the network adapter receives the total packet size per second, the unit is KB/SEC
2, Total-write (-ve), the total packet size of the network adapter sent per second, the unit is KB/SEC
3, Eth0-total, the total size of packets received and sent by the network adapter per second, the unit is KB/SEC

F.sheet Name
1, Sys_summ: System Summary, Blue Line for CPU occupancy change situation, pink for disk IO changes;
2, AAA: about the operating system and nmon itself some information;
3, Bbbp:vmtune, Schedtune, emstat and lsattr command output information;
4. Cpunn: Shows the CPU usage between executions, including user%, sys%, wait% and idle%;
5, Cpu_summ: The occupancy of each CPU in the execution time, which contains user%, sys%, wait% and idle%;
6, Diskbsize: the transmission block size of each hdisk in the execution time;
7, Diskread: Average reading of each hdisk;
8, Diskwrite: The average write situation of each hdisk;
9, Diskxfer: Every hdisk I/O per second operation;
10. Jfsfile: This sheet shows the percentage of space that is being used in each interval for each file system
11, Netpacket: This sheet statistics the number of read and write packets per adapter network
12. PROC: This sheet contains statistical information inside the Nmon kernel. Where the Runqueue and swap-in fields are the average time interval used, the units of the other items are ratios per second
13, ZZZZ: This sheet automatic conversion of all Nmon time stamp for the current real time, easy to analyze

Linux System performance monitoring tool Nmon