Use the performance analysis tool nmon in linux
1. Introduction The nmon tool can help display all important performance optimization information on a single screen and dynamically update it. This efficient tool can work on any dumb screen, telnet session, or even dial-up line. In addition, it does not consume a lot of CPU cycles, usually less than 2%. On the updated computer, the CPU usage will be lower than 1%. Use a dummy screen to display data on the screen and update the data every two seconds. However, you can easily change the time interval to a longer or shorter time period. If you stretch the window and display the data in Windows X, 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 reports. The output file is in the format of a workbook (.csv ). The nmon tool allows AIX and Linux performance experts to monitor and analyze performance data, including:
- CPU usage
- Memory usage
- Kernel statistics and running queue Information
- Disk I/O speed, transfer and read/write Ratio
- Available space in the file system
- Disk Adapter
- Network I/O speed, transmission and read/write Ratio
- Page space and page speed
- CPU and AIX specifications
- Processes that consume the most resources
- Ibm http Web Cache
- Custom disk groups
- Computer details and resources
- Asynchronous I/O, only applicable to AIX
- Workload Manager (WLM), applicable only to AIX
- IBM TotalStorage? Enterprise Storage Server? (ESS) disk, only applicable to AIX
- Network File System (NFS)
- Dynamic LPAR (DLPAR) changes are only applicable to pSeries p5 and OpenPower for AIX or Linux. They also include a graph generated from nmon output and can be displayed on the Web site. A new tool for gif files.
2 get the http://sourceforge.jp/projects/sfnet_nmon/releases/ or get it directly from this, it also contains the analysis tool
3. After decompression, we can see the files on each platform. We only need to use the appropriate files, generally nmon_linux_x86_64. $ Cp nmon_linux_x86_64 nmon $. /When nmon is used for the first time, the system prompts ". /nmon: Permission denied "run chnod 777. /nmon, you can see the specific usage on this interface. For example, if you enter "n", the network I/O and 'D' are displayed, and the disk information is displayed, 'M' displays memory information, and so on;
4. For other options, we generally use-s and-c-f options, for example, $. /nmon-s 1-c 3600-f: the output frequency is 1 second, 3600 times in total, that is, 1 hour. The file name is saved as 5. The above toolkit has included nmon analyser v33g.xls, open the 'options' option and select fried ky "http://www.bkjia.com/kf/web/php/" target = "_ blank" class = "keylink"> PHP1Mq + tMvE2sjdo6zIu7rzteO79w = "analyze nmon notebook, which contains detailed reports.
After the six nmon report analysis report is generated, various data and detailed table data will be generated, including network, disk, CPU, etc. For detailed parameters and descriptions, refer to the NET option as an example, this report is based on a lightweight network library. It uses the epoll model and has 10000 connections. Each connection is sent at 10 packets/sec frequency, and each packet is about 100 bytes in size, the Code comes from the https://github.com/yuyunliuhen/tinynet. Monitoring is divided into two sections, the first section is the first hour of connection until all connections are successful, and the second section is the one hour after the stable operation for 12 hours. Section 1: Starting from, network traffic increases linearly as the number of connections increases. At, network traffic tends to be stable. Because it is an echo model, the read and write operations are basically the same, stable mean after 10 M/se 127y "http://www.bkjia.com/kf/yidong/wp/" target = "_ blank" class = "keylink"> WPX89PSoaMKtdq2/rboo7okpgltzybzcm9 "http://www.2cto.com/uploadfile/Collfiles/20141202/20141202092353158.png" alt = "\"> traffic mean or around 10 M/sec, the maximum read value is 50 M/sec, and the maximum write value is 38 M/sec. For 1000 Mb NICs, 10 Mb is much lower than 125 Mb/8 = Mb, and there is no bottleneck on bandwidth traffic. Other performance data charts are similar. We will not list them here! Reference: http://baike.baidu.com/link? Url = UYp8YULTXT04CLEHGP0HsYFff_5k1m-lyDGLFgGi_UDWSxu53T9lpo8YL4Fo6iLnQ-1_olo-M3UfIVR1E4wLGa
