Open the Nmon result file with analysis tools such as Nmon_analyser_hzt.xls, and if prompted with the inability to load macros, click Tools - macro - security, secure and low, Once saved, open it again.
? Sys_summ page, Rollup for server resource Usage
The main data we need are CPU,mem,io and net. such as:
Fill in the test report with the above contents truthfully.
? Cpu_all page, which is The detailed data for the server CPU utilization
Our main focus is on:
Cpu usage is divided into three parts, the system, the user, and the wait, respectively, corresponding to the red, blue, and green portions of the graph. CPU Utilization 100% , if the red accounted for the majority, the system process itself occupies most of the CPU Resources, Consider whether there are too many zombie processes in the system or if the system process has a dead loop, etc., if the majority of the blue, the user process consumes a lot of CPU Resources, you can consider whether the pressure system is too heavy, or the pressure system has a large number of operations such as consumption CPU the operation of the resource.
? Mem page, which is an overview of server memory utilization.
The blue lines in the figure are physical memory usage, and the red lines are virtual memory usage. We counted the memory usage as: (used physical memory + used virtual memory)/(physical memory + virtual memory)x100% .
? Memnew page, which is the detailed data for server memory utilization
We focus on:
The blue part of the figure is the system consumes memory, the purple is the process consumes the memory, the yellow is the file occupies the memory. Assuming that the memory utilization 100% , such as blue + purple portion is low, the rest is all yellow, then the system has a large number of file operations, the file is written to memory to speed up read and write, does not affect the actual use of memory efficiency; + purple portion accounted for more than 90%, it indicates that memory usage is high, need to combine page page to analyze whether there are memory leaks and so on.
? page, which is the use of swap space
Pgin number of pages read per second, including the number of pages read from the file system
Pgout pages written per second, including the number of pages written to the filesystem
Pgsin number of pages read per second from the page space
Pgsout number of pages written to page space per second
The data computed by the Fsin Analyzer is used for pgin-pgsin Graphics processing
The data computed by the Fsout Analyzer is used for pgout-pgsout Graphics processing
If there is a large number of frequent swap-out phenomena, there is a frequent switchover between physical memory and virtual memory, and physical memory may be problematic.
? Disk_summ page, disk
Disk read The number of kilobytes read per second in the blue section of the graph
Disk Write The number of kilobytes written per second in the red part of the graph
Io/sec the number of io per second (once io is the control operation one time read or write,io block is the size of the read or write), This is the black line above, the value of this parameter is written in the report .
? NET page Network
Total read The number of kilobytes received per second, blue part
Total write the number of kilobytes sent per second, the red part
Network =total Read-total write
Network indicators are generally based on the equipment to determine
? Top page, which is the top event for various usage rates
For example, thetop page records CPU usage per process,mem usage,disk usage, and Net usage rate. We can easily crawl to the most CPU -intensive or most mem -intensive programs, providing clues for tuning.
Nmon Results Analysis