ActiveMQ極限性能參數測試結果-Jmeter

ActiveMQ極限性能參數測試結果

http://blog.csdn.net/geniuslb2001/article/details/8426025

http://jm.taobao.org/2016/04/01/kafka-vs-rabbitmq-vs-rocketmq-message-send-performance/

Apache JMeter™

http://jmeter.apache.org/

Jmeter下載路徑

http://jmeter.apache.org/download_jmeter.cgi

下載Binaries文件

apache-jmeter-3.0.tgz	md5	sha	pgp
apache-jmeter-3.0.zip	md5	sha	pgp

JMeter使用安裝使用方法，參考博文：

http://jmeter.apache.org/usermanual/get-started.html#libraries_activemq

To run JMeter, run the jmeter.bat (for Windows) or jmeter (for Unix) file. These files are found in the bin directory. After a short time, the JMeter GUI should appear.

Building a JMS Point-to-Point Test Plan

http://jmeter.apache.org/usermanual/build-jms-point-to-point-test-plan.html

jmeter參考例子

http://blog.jassassin.com/2014/04/17/tools/jmeter/

jmeter配置成功activemq Point to Point 運行例子

運行步驟一、

apache-jmeter-3.0\bin\ApacheJMeter.jar

將 jndi.properties 文件添加到ApacheJMeter.jar裏面。

java.naming.factory.initial = org.apache.activemq.jndi.ActiveMQInitialContextFactory  
java.naming.provider.url = failover://(tcp://test3-dcb:61616,tcp://test2-dcb:61616,tcp://test1:61616)##這裏能夠是tcp或者是failover協議  
(QueueConnectionFactory)context.lookup("connectionFactory") 
connectionFactoryNames = connectionFactory  
(Queue)context.lookup("MyQueue")
queue.MyQueue = example.MyQueue

在這裏配置和步驟二中配置的 your_script.jmx能夠起到同樣的效果。

運行步驟2、

添加sampleer

http://blog.csdn.net/hunterno4/article/details/10161081#

jmeter non gui 使用

使用命令：jmeter -n -t your_script.jmx

例子：

<hashTree>
      <ThreadGroup guiclass="ThreadGroupGui" testclass="ThreadGroup" testname="activemqTesting" enabled="true">
        <stringProp name="ThreadGroup.on_sample_error">continue</stringProp>
        <elementProp name="ThreadGroup.main_controller" elementType="LoopController" guiclass="LoopControlPanel" testclass="LoopController" testname="Loop Controller" enabled="true">
          <boolProp name="LoopController.continue_forever">false</boolProp>
          <intProp name="LoopController.loops">-1</intProp>
        </elementProp>
        <stringProp name="ThreadGroup.num_threads">100</stringProp>
        <stringProp name="ThreadGroup.ramp_time">0</stringProp>
        <longProp name="ThreadGroup.start_time">1477399580000</longProp>
        <longProp name="ThreadGroup.end_time">1477399580000</longProp>
        <boolProp name="ThreadGroup.scheduler">false</boolProp>
        <stringProp name="ThreadGroup.duration"></stringProp>
        <stringProp name="ThreadGroup.delay"></stringProp>
      </ThreadGroup>
      <hashTree>
        <JMSSampler guiclass="JMSSamplerGui" testclass="JMSSampler" testname="activemq-queue" enabled="true">
          <stringProp name="JMSSampler.queueconnectionfactory">connectionFactory</stringProp>
          <stringProp name="JMSSampler.SendQueue">MyQueue</stringProp>
          <stringProp name="JMSSampler.ReceiveQueue">MyQueue</stringProp>
          <boolProp name="JMSSampler.isFireAndForget">true</boolProp>
          <boolProp name="JMSSampler.isNonPersistent">false</boolProp>
          <boolProp name="JMSSampler.useReqMsgIdAsCorrelId">false</boolProp>
          <stringProp name="JMSSampler.timeout"></stringProp>
          <stringProp name="HTTPSamper.xml_data"></stringProp>
          <stringProp name="JMSSampler.initialContextFactory">org.apache.activemq.jndi.ActiveMQInitialContextFactory</stringProp>
          <stringProp name="JMSSampler.contextProviderUrl">failover://(tcp://test3-dcb:61616,tcp://test2-dcb:61616,tcp://test1:61616)</stringProp> ####failover://(tcp://test3-dcb:61616,tcp://test2-dcb:61616,tcp://test1:61616)
          <elementProp name="JMSSampler.jndiProperties" elementType="Arguments" guiclass="ArgumentsPanel" testclass="Arguments" testname="User Defined Variables" enabled="true">
            <collectionProp name="Arguments.arguments"/>
          </elementProp>
          <elementProp name="arguments" elementType="JMSProperties">
            <collectionProp name="JMSProperties.properties">
              <elementProp name="" elementType="JMSProperty">
                <stringProp name="JMSProperty.name">obj_name</stringProp>
                <stringProp name="JMSProperty.value">final11111</stringProp>
                <stringProp name="JMSProperty.type">java.lang.String</stringProp>
              </elementProp>
              <elementProp name="" elementType="JMSProperty">
                <stringProp name="JMSProperty.name">obj_type</stringProp>
                <stringProp name="JMSProperty.value">/service</stringProp>
                <stringProp name="JMSProperty.type">java.lang.String</stringProp>
              </elementProp>
              <elementProp name="" elementType="JMSProperty">
                <stringProp name="JMSProperty.name">id</stringProp>
                <stringProp name="JMSProperty.value">c064a027-ba94-460b-8831-8d8031fc686c</stringProp>
                <stringProp name="JMSProperty.type">java.lang.String</stringProp>
              </elementProp>
              <elementProp name="" elementType="JMSProperty">
                <stringProp name="JMSProperty.name">created_t</stringProp>
                <stringProp name="JMSProperty.value">1477472363</stringProp>
                <stringProp name="JMSProperty.type">java.lang.String</stringProp>
              </elementProp>
              <elementProp name="" elementType="JMSProperty">
                <stringProp name="JMSProperty.name">old_obj_name</stringProp>
                <stringProp name="JMSProperty.value"></stringProp>
                <stringProp name="JMSProperty.type">java.lang.String</stringProp>
              </elementProp>
            </collectionProp>
          </elementProp>
        </JMSSampler>
        <hashTree/>
      </hashTree>

https://www.blazemeter.com/blog/dear-abby-blazemeter-how-do-i-run-jmeter-non-gui-mode?utm_source=BM&utm_medium=BM_blog&utm_campaign=5_Ways_to_Launch_blog

具體細節能夠參考

https://www.blazemeter.com/blog/dear-abby-blazemeter-how-do-i-run-jmeter-non-gui-mode?utm_source=BM&utm_medium=BM_blog&utm_campaign=5_Ways_to_Launch_blog

Dear Abby BlazeMeter, How Do I Run JMeter in Non-GUI Mode?

Do you think the JMeter GUI is a good GUI?  

After all, the GUI is always considerate, shows up on time, is willing to carry it's share of the load....

On one hand, the GUI makes the program more intuitive; on the other hand, it consumes more resources.

Isn’t better to redirect these resources to the generation of useful load tests?

Yes, it is very good idea. Let's review a few typical cases of running JMeter in non-GUI mode.

Step 1. Prepare the JMeter test

• Open your *.jmx file

• Remove all graphic data writers

They will not work in the non-gui mode and there no need to keep them in testing scenario.

Step 2. Run the JMeter test

• Go to bin folder of your JMeter installation from command prompt

• Type the following command:

jmeter -n -t D:\TestScripts\script.jmx -l D:\TestScripts\scriptresults.jtl

-n [This specifies JMeter is to run in non-gui mode]

-t  [name of JMX file that contains the Test Plan]

-l  [name of JTL file to log sample results to]

-j  [name of JMeter run log file].

Besides these options, JMeter has several other parameters that can be used for running in non-GUI mode.

-R [list of remote servers] Run the test in the specified remote servers

-H [proxy server hostname or ip address]

-P [proxy server port]

These options are used for remote execution of JMeter tests and for using JMeter through a proxy server.

Step 3. Stop running the test

When running JMeter in non-GUI mode, there is no Menu, and JMeter does not react to keystrokes such as Control + '.'. So in versions post 2.3.2, JMeter non-GUI mode will listen for commands on a specific port (default 4445, see the JMeter property jmeterengine.nongui.port ). In versions post 2.4, JMeter supports the automatic choice of an alternate port if the default port is being used (i.e. by another JMeter instance). In this case, JMeter will try the next highest port, continuing until it reaches the JMeter property jmeterengine.nongui.maxport, which defaults to 4455. If maxportis less than or equal to port , port scanning will not take place.

Note that JMeter 2.4 and earlier did not set up the listener for non-GUI clients, only non-GUI standalone tests; this has been fixed.

The chosen port is displayed in the console window.

The commands currently supported are:

Shutdown - graceful shutdown

stoptest  - immediate shutdown

 

These commands can be sent by using the shutdown[.cmd|.sh] or stoptest[.cmd|.sh] script respectively. The scripts are to be found in the JMeter /bin directory. The commands will only be accepted if the script is run from the same host.

Note-before using shutdown.cmd or shutdown.sh CLASSPATH variable should be set to path with ApacheJmeter.jar, because inside these files we have string like

「java -cp %~dp0ApacheJMeter.jar org.apache.jmeter.util.ShutdownClient StopTestNow %*」. It means that JVM calls class file from ApacheJMeter.jar.

Look out for the next episode when we discuss interpreting how the *.jtl files feels about communication and if that is created as a result of non-GUI test.

參考博文：

http://jmeter.apache.org/usermanual/build-test-plan.html

9 Easy Solutions for a JMeter Load Test 「Out of Memory」 Failure

https://www.blazemeter.com/blog/9-easy-solutions-jmeter-load-test-%E2%80%9Cout-memory%E2%80%9D-failure

JMeter Result Analysis

參考http://artoftesting.com/performanceTesting/resultAnaysis.html

After executing a load test we need to interpret the result. For result interpretation we use Listeners that provide different matrices about the load test. In this post we will see the most commonly used matrices, their meanings and the how they are calculated internally from raw data.

Result matrices by Aggregate Report-

Label- Label is the name of the sample or the Transaction Controller

# Samples - The total number of samples corresponding to a given sample

Average - The average time taken (in milliseconds) to execute the requests under a given label. So, if there are 10 samples getting executed then average time taken will be-
Average = Total time taken by all samples /#samples

Median - The median is basically the middle value of response time in the sorted list of samples

90% Line - The Apache JMeter manual describes 90% line as- "90% of the samples took no more than this time". It is actually the 90 percentile of the response times of the samples -
90 percentile = (90/100)*N+1/2 where N is the number of samples
So, if there are 10 samples then 90%line will be 9.5 or 9. It means the 9th value in the sorted list of samples (sorted according to ascending order of their response times) will be the 90% line value.

Min - The minimum time (in milliseconds) taken by the sample

Max - The maximum time (in milliseconds) taken by the sample

Error % - Percentage of errors in the samples

Throughput - Throughput as we all know is output per unit time. In JMeter terms we define throughput as the amount of load applied on the server. So, numerically-
Throughput = Total number of requests to the server/ Total time
                   = Total number of requests to the server/(End time of last sample -Start time of first sample)
Here we just defined Total time in which load was applied on the server as time duration between Start of first sample and end of last sample.

KB/sec - The metric KB/sec is nothing but the throughput measured in terms of bytes. So, KB/sec = (Throughput*Average bytes) /1024 Here Average bytes is the average value of the sample response in bytes and the term 1024 is used to convert the value (Throughput*Average bytes) into kilobytes.