JMeter Pressure test environment preparation

JMeter Performance Monitoring

1. Configure the JMeter environment on the test machine: http://jmeter-plugins.org/downloads/all/website download Two things:   Jmeterplugins-standard-1.1.3.zip , serveragent-2.2.1.zip
2. Standard ZIP package after decompression, put Jmeterplugins-standard.jar under Xxx\apache\jmeter\lib\ext, restart JMeter can see the thread group right-click Listener can see a lot of added content , add the [email protected]-PerfMon Metrics Collector to the existing project as a monitoring server-side data.
3. added [email protected]-PerfMon Metrics collector, in which add Row adds data information that needs to be monitored, such as cpu,memory,disks/ The port defaults to 4444 for information such as IO,TCP.
4. Server-side configuration, unzip the Serveragent.zip package, and place it on the server:

• LINUX:CD command to this folder, enter the command./xxxxxx/startagent.sh. Note that if the client is not configured for Port 4444, the port for the client setting needs to be started, and the command is./startagent.sh--udp-port 0--tcp-port xxxx, of course xxxx is the set port number.
• Windows:, double-click Startagent.bat to start the default port 4444.

5. On the test machine, click Start to start the interface test and stress testing work, in the selected [email protected]-PerfMon Metrics collector can see the performance information from the server.

JMeter Distributed Deployment

The following steps are used to generate the load using multiple machines:

1. Install JMeter on all machines that expect to run JMeter as a load generator, and identify one of the machines as the controller, and other machines as agents.
2. In the JMeter%jmeter_home%/bin directory of the controller machine, locate the Jmeter.properties file and edit the file:

find: remote_hosts=127.0.0.1, modified to: remote_hosts=10.8.39.114:1099,10.8.33.181:1099,localhost:1099

　　Special attention is paid to the port, and some data indicate that port 1644 is the default RMI port number for communication between the Controller and the agent for JMeter, but when tested, it was found that the set to 1644 ran unsuccessfully and was changed to 1099 after it was run. Also pay attention to the agent's machine whether the firewall is turned on and so on.

3. Run the Jmeter-server file on all agent machines

4. Start the JMeter application Jmeter.bat on the controller machine, select the menu "Run"---> "Remote Start" ("Run"---> "startup") to start the agent separately, or you can directly select the " Remote Start All "(" remotely start all ") to start all agents. When

Remote access, you will see a line printed on the console: starting the test on host [ip]:1099 @ ...., the remote execution ends, a line is printed: finished the test on host [ip]:1099 @ ....

Points to note:

1) JMeter Distributed control process, the number of threads initiated by each agent equals the configuration in the thread group, not the configuration in the split thread group, the number of threads that each agent tests individually is the number of threads that you configure on the controller. That means I have 100 threads on the controller, and immediately

Execute, Loop once, then my deployment will have a colleague trigger the number of threads (100X3), execute immediately, execute once. Instead of three machines sharing 100 of threads together.

2) When you want to run the agent on the Contrller machine, you need to execute the Jmeter-server file before executing the Jmeter.bat

3) When configured, ensure that the machine's firewall is turned off.

References:

Performance Monitoring plugin:
Http://www.tuicool.com/articles/RbAN3ea
Performance Monitoring introduction:
http://blog.csdn.net/defonds/article/details/41650813

JMeter Pressure test environment preparation