xdocs/usermanual/remote-test.xml - jmeter - Git at Google

 <?xml version="1.0"?>
 <!--
    Licensed to the Apache Software Foundation (ASF) under one or more
    contributor license agreements.  See the NOTICE file distributed with
    this work for additional information regarding copyright ownership.
    The ASF licenses this file to You under the Apache License, Version 2.0
    (the "License"); you may not use this file except in compliance with
    the License.  You may obtain a copy of the License at

        http://www.apache.org/licenses/LICENSE-2.0

    Unless required by applicable law or agreed to in writing, software
    distributed under the License is distributed on an "AS IS" BASIS,
    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and
    limitations under the License.
 -->

 <!DOCTYPE document[
 <!ENTITY sect-num '15'>
 <!ENTITY hellip   "&#x02026;" >
 ]>

 <document prev="listeners.html" next="generating-dashboard.html" id="$Id$">

 <properties>
   <title>User's Manual: Remote (Distributed) Testing</title>
 </properties>

 <body>

 <section name="&sect-num;. Remote Testing">

 <p>In the event that your JMeter client machine is unable, performance-wise, to simulate
 enough users to stress your server or is limited at network level, an option exists to control multiple, remote JMeter
 engines from a single JMeter client.  By running JMeter remotely, you can replicate
 a test across many low-end computers and thus simulate a larger load on the server.  One
 instance of the JMeter client can control any number of remote JMeter instances, and collect
 all the data from them.  This offers the following features:

 <ul>
 <li>Saving of test samples to the local machine </li>
 <li>Management of multiple JMeterEngines from a single machine </li>
 <li>No need to copy the test plan to each server - the client sends it to all the servers</li>
 </ul>
 </p>
 <note>
 Note: The same test plan is run by all the servers.
 JMeter does not distribute the load between servers, each runs the full test plan.
 So if you set 1000 Threads and have 6 JMeter server, you end up injecting 6000 Threads.
 </note>
 <p>
 However, remote mode does use more resources than running the same number of non-GUI tests independently.
 If many server instances are used, the client JMeter can become overloaded, as can the client network connection.
 This has been improved by switching to Stripped modes (see below) but you should always check that your client is not overloaded.
 </p>
 <p>Note that while you can execute the JMeterEngine on your application
 server, you need to be mindful of the fact that this will be adding processing
 overhead on the application server and thus your testing results will be
 somewhat tainted.  The recommended approach is to have one or more machines on
 the same Ethernet segment as your application server that you configure to run
 the JMeter Engine.  This will minimize the impact of the network on the test
 results without impacting the performance of the application server
 itself.
 </p>

 <p><b>Step 0: Configure the nodes</b></p>
 <p>
 Make sure that all the nodes (client and servers) :
 <ul>
     <li>are running exactly the same version of JMeter.</li>
     <li>are using the same version of Java on all systems. Using different versions of Java may work but is discouraged.</li>
 </ul>
 </p>
 <p>
 If the test uses any data files, <b>note that these are not sent across by the client so
 make sure that these are available in the appropriate directory on each server</b>.
 If necessary you can define different values for properties by editing the <code>user.properties</code> or <code>system.properties</code>
 files on each server. These properties will be picked up when the server is started and may be
 used in the test plan to affect its behaviour (e.g. connecting to a different remote server).
 Alternatively use different content in any datafiles used by the test
 (e.g. if each server must use unique ids, divide these between the data files)
 </p>

 <p><b>Step 1: Start the servers </b></p>
 <p>To run JMeter in remote node, start the JMeter server component on all machines you wish to run on by running
 the <code>JMETER_HOME/bin/jmeter-server</code> (unix) or <code>JMETER_HOME/bin/jmeter-server.bat</code> (windows) script.</p>
 <p>Note that there can only be one JMeter server on each node unless different RMI ports are used.</p>
 <p>Since JMeter 2.3.1, the JMeter server application starts the RMI registry itself;
 there is no need to start RMI registry separately.
 To revert to the previous behaviour, define the JMeter property <source>server.rmi.create=false</source> on the server host systems.
 </p>
 <p>
 By default, RMI uses a dynamic port for the JMeter server engine. This can cause problems for firewalls,
 so with you can define the JMeter property <code>server.rmi.localport</code>
 to control this port number.
 If this is non-zero, it will be used as the local port number for the server engine.
 </p>
 <p><b>Step 2: Add the server IP to your client's Properties File</b></p>
 <p>Edit the properties file <i>on the controlling JMeter machine</i>.  In <code>JMETER_HOME/bin/jmeter.properties</code>,
 find the property named, "<code>remote_hosts</code>", and
 add the value of your running JMeter server's IP address.  Multiple such servers can be added, comma-delimited.</p>
 <p>Note that you can use the <code>-R</code> <a href="get-started.html#override">command line option</a>
 instead to specify the remote host(s) to use. This has the same effect as using <code>-r</code> and <code>-Jremote_hosts={serverlist}</code>.
     E.g. </p><source>jmeter -Rhost1,127.0.0.1,host2</source>
 <p>If you define the JMeter property <code>server.exitaftertest=true</code>, then the server will exit after it runs a single test.
 See also the <code>-X</code> flag (described below)
 </p>
 <p><b>Step 3a: Start the JMeter Client from a GUI client to check configuration</b></p>
 <p>Now you are ready to start the controlling JMeter client. For MS-Windows, start the client with the script "<code>bin/jmeter.bat</code>".  For UNIX,
 use the script "<code>bin/jmeter</code>".  You will notice that the Run menu contains two new sub-menus: "Remote Start" and "Remote Stop"
 (see figure 1). These menus contain the client that you set in the properties file.  Use the remote start and stop instead of the
 normal JMeter start and stop menu items.</p>
 <figure image="remote/run-menu00.png" width="487" height="295">Figure 1 - Run Menu</figure>

 <p><b>Step 3b: Start the JMeter from a non-GUI Client</b></p>
 <p>
 GUI mode should only be used for debugging, as a better alternative, you should start the test on remote server(s) from a non-GUI (command-line) client.
 The command to do this is:</p>
 <source>
 jmeter -n -t script.jmx -r
 </source>
 or
 <source>
 jmeter -n -t script.jmx -R server1,server2,&hellip;
 </source>
 Other flags that may be useful:
 <dl>
 <dt><code>-Gproperty=value</code></dt><dd>define a property in all the servers (may appear more than once)</dd>
 <dt><code>-X</code></dt><dd>Exit remote servers at the end of the test.</dd>
 </dl>
 The first example will start the test on whatever servers are defined in the JMeter property <code>remote_hosts</code>;<br/>
 The second example will define <code>remote_hosts</code> from the list of servers and then start the test on the remote servers.
 <br/>
 The command-line client will exit when all the remote servers have stopped.

 <subsection name="&sect-num;.1 Doing it Manually" anchor="detail_instructions">
 <p>In some cases, the jmeter-server script may not work for you (if you are using an OS platform not anticipated by the JMeter developers).
 Here is how to start the JMeter servers (step 1 above) with a more manual process:</p>
 <p><b>Step 1a: Start the RMI Registry</b></p>
 <p>
 Since JMeter 2.3.1, the RMI registry is started by the JMeter server, so this section does not apply in the normal case.
 To revert to the previous behaviour, define the JMeter property <code>server.rmi.create=false</code> on the server host systems
 and follow the instructions below.
 </p>
 <p>JMeter uses Remote Method Invocation (RMI) as the remote communication mechanism.  Therefore, you need
 to run the RMI Registry application (which is named, "<code>rmiregistry</code>") that comes with the JDK and is located in the "<code>bin</code>"
 directory.  Before running <code>rmiregistry</code>, make sure that the following jars are in your system classpath:
 <ul>
     <li><code>JMETER_HOME/lib/ext/ApacheJMeter_core.jar</code></li>
     <li><code>JMETER_HOME/lib/jorphan.jar</code></li>
     <li><code>JMETER_HOME/lib/logkit-2.0.jar</code></li>
 </ul>
 The
 rmiregistry application needs access to certain JMeter classes.  Run <code>rmiregistry</code> with no parameters.  By default the
 application listens to port <code>1099</code>.</p>

 <p><b>Step 1b: Start the JMeter Server</b></p>
 <p>Once the RMI Registry application is running, start the JMeter Server.
 Use the "<code>-s</code>" option with the jmeter startup script ("<code>jmeter -s</code>").</p>

 <p>Steps 2 and 3 remain the same.</p>
 </subsection>
 <subsection name="&sect-num;.2 Tips" anchor="tips">
 <p>
 JMeter/RMI requires a connection from the client to the server. This will use the port you chose, default <code>1099</code>.<br/>
 JMeter/RMI also requires a reverse connection in order to return sample results from the server to the client.<br/>
 This will use a high-numbered port. <br/>
 This port can be controlled by jmeter property called <code>client.rmi.localport</code> in <code>jmeter.properties</code>.<br/>
 If there are any firewalls or other network filters between JMeter client and server,
 you will need to make sure that they are set up to allow the connections through.
 If necessary, use monitoring software to show what traffic is being generated.
 </p>
 <p>If you're running Suse Linux, these tips may help. The default installation may enable the firewall. In that case,
 remote testing will not work properly. The following tips were contributed by Sergey Ten.</p>
 <p>If you see connections refused, turn on debugging by passing the following options.</p>
 <source>
 rmiregistry -J-Dsun.rmi.log.debug=true \
      -J-Dsun.rmi.server.exceptionTrace=true \
      -J-Dsun.rmi.loader.logLevel=verbose \
      -J-Dsun.rmi.dgc.logLevel=verbose \
      -J-Dsun.rmi.transport.logLevel=verbose \
      -J-Dsun.rmi.transport.tcp.logLevel=verbose \
 </source>
 <p>Since JMeter 2.3.1, the RMI registry is started by the server; however the options can still be passed in from the JMeter command line.
 For example: "<code>jmeter -s -Dsun.rmi.loader.logLevel=verbose</code>" (i.e. omit the <code>-J</code> prefixes).
 Alternatively the properties can be defined in the <code>system.properties</code> file.
 </p>
 <p>The solution to the problem is to remove the loopbacks <code>127.0.0.1</code> and <code>127.0.0.2</code> from <code>/etc/hosts</code>.
 What happens is <code>jmeter-server</code> can't connect to rmiregistry if <code>127.0.0.2</code> loopback is not available.
 Use the following settings to fix the problem.</p>
 <p> Replace</p>
 <source>`dirname $0`/jmeter  -s "$@"</source>
 <p> With</p>
 <source>
 HOST="-Djava.rmi.server.hostname=[computer_name][computer_domain] \
       -Djava.security.policy=`dirname $0`/[policy_file]" \
 `dirname $0`/jmeter $HOST -s "$@"
 </source>
 <p>Also create a policy file and add <code>[computer_name][computer_domain]</code> line to <code>/etc/hosts</code>.</p>

 <p>In order to better support SSH-tunneling of the RMI communication channels used
 in remote testing, since JMeter 2.6:</p>
 <ul>
 <li>a new property "<code>client.rmi.localport</code>" can be set to control the RMI port used by the RemoteSampleListenerImpl</li>
 <li>To support tunneling RMI traffic over an SSH tunnel as the remote endpoint using a port on the local machine,
  loopback interface is now allowed to be used if it has been specified directly using the Java System Property
  "<code>java.rmi.server.hostname</code>" parameter.</li>
 </ul>
 </subsection>
 <subsection name="&sect-num;.3 Using a different port" anchor="portchange">
 <p>By default, JMeter uses the standard RMI port <code>1099</code>. It is possible to change this. For this to work successfully,
 all the following need to agree:</p>
 <ul>
 <li>On the server, start <code>rmiregistry</code> using the new port number</li>
 <li>On the server, start JMeter with the property <code>server_port</code> defined</li>
 <li>On the client, update the <code>remote_hosts</code> property to include the new remote <code>host:port</code> settings</li>
 </ul>

 <p>Since Jmeter 2.1.1, the jmeter-server scripts provide support for changing the port.
 For example, assume you want to use port <code>1664</code> (perhaps <code>1099</code> is already used).</p>
 On Windows (in a DOS box)
 <source>
 C:\JMETER> SET SERVER_PORT=1664
 C:\JMETER> JMETER-SERVER [other options]
 </source>
 On Unix:
 <source>
 $ SERVER_PORT=1664 jmeter-server [other options]
 </source>
 [N.B. use upper case for the environment variable]
 <p>
 In both cases, the script starts rmiregistry on the specified port,
 and then starts JMeter in server mode, having defined the "<code>server_port</code>" property.
 </p>
 <p>
 The chosen port will be logged in the server <code>jmeter.log</code> file (<code>rmiregistry</code> does not create a log file).
 </p>
 </subsection>

 <subsection name="&sect-num;.4 Using a different sample sender" anchor="sendermode">
 <p>
 Listeners in the test plan send their results back to the client JMeter which writes the results to the specified files
 By default, samples are sent back synchronously as they are generated.
 This can affect the maximum throughput of the server test; the sample result has to be sent back before the thread can
 continue.
 There are some JMeter properties that can be set to alter this behaviour.
 </p>
 <dl>
 <dt><code>mode</code></dt><dd>sample sending mode - default is <code>StrippedBatch</code> since 2.9. This should be set on the client node.
     <dl>
     <dt><code>Standard</code></dt><dd>send samples synchronously as soon as they are generated</dd>
     <dt><code>Hold</code></dt><dd>hold samples in an array until the end of a run. This may use a lot of memory on the server and is discouraged.</dd>
     <dt><code>DiskStore</code></dt><dd>store samples in a disk file (under <code>java.io.temp</code>) until the end of a run.
     The serialised data file is deleted on JVM exit.</dd>
     <dt><code>StrippedDiskStore</code></dt><dd>remove responseData from successful samples, and use DiskStore sender to send them.</dd>
     <dt><code>Batch</code></dt><dd>send saved samples when either the count (<code>num_sample_threshold</code>) or time (<code>time_threshold</code>) exceeds a threshold,
     at which point the samples are sent synchronously.
     The thresholds can be configured on the server using the following properties:
         <dl>
         <dt><code>num_sample_threshold</code></dt><dd>number of samples to accumulate, default <code>100</code></dd>
         <dt><code>time_threshold</code></dt><dd>time threshold, default 60000 ms = 60 seconds</dd>
         </dl>

      See also the Asynch mode, described below.</dd>
     <dt><code>Statistical</code></dt><dd>send a summary sample when either the count or time exceeds a threshold.
     The samples are summarised by thread group name and sample label.
     The following fields are accumulated:
       <ul>
       <li><code>elapsed time</code></li>
       <li><code>latency</code></li>
       <li><code>bytes</code></li>
       <li><code>sample count</code></li>
       <li><code>error count</code></li>
       </ul>
     Other fields that vary between samples are lost.
     </dd>
     <dt><code>Stripped</code></dt><dd>remove responseData from successful samples</dd>
     <dt><code>StrippedBatch</code></dt><dd>remove responseData from successful samples, and use Batch sender to send them.</dd>
     <dt><code>Asynch</code></dt><dd>samples are temporarily stored in a local queue. A separate worker thread sends the samples.
         This allows the test thread to continue without waiting for the result to be sent back to the client.
         However, if samples are being created faster than they can be sent, the queue will eventually fill up,
         and the sampler thread will block until some samples can be drained from the queue.
         This mode is useful for smoothing out peaks in sample generation.
         The queue size can be adjusted by setting the JMeter property
         <code>asynch.batch.queue.size</code> (default <code>100</code>) on the server node.
     </dd>
     <dt><code>StrippedAsynch</code></dt><dd>remove responseData from successful samples, and use Async sender to send them.</dd>
     <dt><code>Custom implementation</code></dt><dd>set the mode parameter to your custom sample sender class name.
     This must implement the interface <code>SampleSender</code> and have a constructor which takes a single
     parameter of type <code>RemoteSampleListener</code>.
     </dd>
     </dl>
     </dd>
 </dl>
 <note><code>Stripped</code> mode family strips <code>responseData</code> so this means that some Elements that rely
 on the previous <code>responseData</code> being available will not work.<br/>
 This is not really a problem as there is always a more efficient way to implement this feature.
 </note>
 <p>The following properties apply to the <code>Batch</code> and <code>Statistical</code> modes:</p>
     <dl>
     <dt><code>num_sample_threshold</code></dt><dd>number of samples in a batch (default <code>100</code>)</dd>
     <dt><code>time_threshold</code></dt><dd>number of milliseconds to wait (default 60 seconds)</dd>
     </dl>
 </subsection>


 <subsection name="&sect-num;.5 Dealing with nodes that failed starting" anchor="retries">
   <p>
     For large-scale tests there is a chance that some part of remote servers will be unavailable or down.
     For example, when you use automation script to allocate many cloud machines and use them as generators,
     some of requested machines might fail booting because of cloud's issues.
     Since JMeter 2.13 there are new properties to control this behaviour.
   </p>
   <p>
     First what you might want is to retry initialization attempts in hope that failed nodes just slightly delayed their boot.
     To enable retries, you should set <code>client.tries</code> property to total number of connection attempts.
     By default it does only one attempt. To control retry delay, set the <code>client.retries_delay</code> property
     to number of milliseconds to sleep between attempts.
   </p>
   <p>
     Finally, you might still want to run the test with those generators that succeeded initialization and skipping failed nodes.
     To enable that, set the <code>client.continue_on_fail=true</code> property.
   </p>
 </subsection>

 </section>

 </body>
 </document>
	<?xml version="1.0"?>
	<!--
	Licensed to the Apache Software Foundation (ASF) under one or more
	contributor license agreements. See the NOTICE file distributed with
	this work for additional information regarding copyright ownership.
	The ASF licenses this file to You under the Apache License, Version 2.0
	(the "License"); you may not use this file except in compliance with
	the License. You may obtain a copy of the License at

	http://www.apache.org/licenses/LICENSE-2.0

	Unless required by applicable law or agreed to in writing, software
	distributed under the License is distributed on an "AS IS" BASIS,
	WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
	See the License for the specific language governing permissions and
	limitations under the License.
	-->

	<!DOCTYPE document[
	<!ENTITY sect-num '15'>
	<!ENTITY hellip "…" >
	]>

	<document prev="listeners.html" next="generating-dashboard.html" id="$Id$">

	<properties>
	<title>User's Manual: Remote (Distributed) Testing</title>
	</properties>

	<body>

	<section name="&sect-num;. Remote Testing">

	<p>In the event that your JMeter client machine is unable, performance-wise, to simulate
	enough users to stress your server or is limited at network level, an option exists to control multiple, remote JMeter
	engines from a single JMeter client. By running JMeter remotely, you can replicate
	a test across many low-end computers and thus simulate a larger load on the server. One
	instance of the JMeter client can control any number of remote JMeter instances, and collect
	all the data from them. This offers the following features:

	<ul>
	<li>Saving of test samples to the local machine </li>
	<li>Management of multiple JMeterEngines from a single machine </li>
	<li>No need to copy the test plan to each server - the client sends it to all the servers</li>
	</ul>
	</p>
	<note>
	Note: The same test plan is run by all the servers.
	JMeter does not distribute the load between servers, each runs the full test plan.
	So if you set 1000 Threads and have 6 JMeter server, you end up injecting 6000 Threads.
	</note>
	<p>
	However, remote mode does use more resources than running the same number of non-GUI tests independently.
	If many server instances are used, the client JMeter can become overloaded, as can the client network connection.
	This has been improved by switching to Stripped modes (see below) but you should always check that your client is not overloaded.
	</p>
	<p>Note that while you can execute the JMeterEngine on your application
	server, you need to be mindful of the fact that this will be adding processing
	overhead on the application server and thus your testing results will be
	somewhat tainted. The recommended approach is to have one or more machines on
	the same Ethernet segment as your application server that you configure to run
	the JMeter Engine. This will minimize the impact of the network on the test
	results without impacting the performance of the application server
	itself.
	</p>

	<p><b>Step 0: Configure the nodes</b></p>
	<p>
	Make sure that all the nodes (client and servers) :
	<ul>
	<li>are running exactly the same version of JMeter.</li>
	<li>are using the same version of Java on all systems. Using different versions of Java may work but is discouraged.</li>
	</ul>
	</p>
	<p>
	If the test uses any data files, <b>note that these are not sent across by the client so
	make sure that these are available in the appropriate directory on each server</b>.
	If necessary you can define different values for properties by editing the <code>user.properties</code> or <code>system.properties</code>
	files on each server. These properties will be picked up when the server is started and may be
	used in the test plan to affect its behaviour (e.g. connecting to a different remote server).
	Alternatively use different content in any datafiles used by the test
	(e.g. if each server must use unique ids, divide these between the data files)
	</p>

	<p><b>Step 1: Start the servers </b></p>
	<p>To run JMeter in remote node, start the JMeter server component on all machines you wish to run on by running
	the <code>JMETER_HOME/bin/jmeter-server</code> (unix) or <code>JMETER_HOME/bin/jmeter-server.bat</code> (windows) script.</p>
	<p>Note that there can only be one JMeter server on each node unless different RMI ports are used.</p>
	<p>Since JMeter 2.3.1, the JMeter server application starts the RMI registry itself;
	there is no need to start RMI registry separately.
	To revert to the previous behaviour, define the JMeter property <source>server.rmi.create=false</source> on the server host systems.
	</p>
	<p>
	By default, RMI uses a dynamic port for the JMeter server engine. This can cause problems for firewalls,
	so with you can define the JMeter property <code>server.rmi.localport</code>
	to control this port number.
	If this is non-zero, it will be used as the local port number for the server engine.
	</p>
	<p><b>Step 2: Add the server IP to your client's Properties File</b></p>
	<p>Edit the properties file <i>on the controlling JMeter machine</i>. In <code>JMETER_HOME/bin/jmeter.properties</code>,
	find the property named, "<code>remote_hosts</code>", and
	add the value of your running JMeter server's IP address. Multiple such servers can be added, comma-delimited.</p>
	<p>Note that you can use the <code>-R</code> <a href="get-started.html#override">command line option</a>
	instead to specify the remote host(s) to use. This has the same effect as using <code>-r</code> and <code>-Jremote_hosts={serverlist}</code>.
	E.g. </p><source>jmeter -Rhost1,127.0.0.1,host2</source>
	<p>If you define the JMeter property <code>server.exitaftertest=true</code>, then the server will exit after it runs a single test.
	See also the <code>-X</code> flag (described below)
	</p>
	<p><b>Step 3a: Start the JMeter Client from a GUI client to check configuration</b></p>
	<p>Now you are ready to start the controlling JMeter client. For MS-Windows, start the client with the script "<code>bin/jmeter.bat</code>". For UNIX,
	use the script "<code>bin/jmeter</code>". You will notice that the Run menu contains two new sub-menus: "Remote Start" and "Remote Stop"
	(see figure 1). These menus contain the client that you set in the properties file. Use the remote start and stop instead of the
	normal JMeter start and stop menu items.</p>
	<figure image="remote/run-menu00.png" width="487" height="295">Figure 1 - Run Menu</figure>

	<p><b>Step 3b: Start the JMeter from a non-GUI Client</b></p>
	<p>
	GUI mode should only be used for debugging, as a better alternative, you should start the test on remote server(s) from a non-GUI (command-line) client.
	The command to do this is:</p>
	<source>
	jmeter -n -t script.jmx -r
	</source>
	or
	<source>
	jmeter -n -t script.jmx -R server1,server2,…
	</source>
	Other flags that may be useful:
	<dl>
	<dt><code>-Gproperty=value</code></dt><dd>define a property in all the servers (may appear more than once)</dd>
	<dt><code>-X</code></dt><dd>Exit remote servers at the end of the test.</dd>
	</dl>
	The first example will start the test on whatever servers are defined in the JMeter property <code>remote_hosts</code>;<br/>
	The second example will define <code>remote_hosts</code> from the list of servers and then start the test on the remote servers.
	<br/>
	The command-line client will exit when all the remote servers have stopped.

	<subsection name="&sect-num;.1 Doing it Manually" anchor="detail_instructions">
	<p>In some cases, the jmeter-server script may not work for you (if you are using an OS platform not anticipated by the JMeter developers).
	Here is how to start the JMeter servers (step 1 above) with a more manual process:</p>
	<p><b>Step 1a: Start the RMI Registry</b></p>
	<p>
	Since JMeter 2.3.1, the RMI registry is started by the JMeter server, so this section does not apply in the normal case.
	To revert to the previous behaviour, define the JMeter property <code>server.rmi.create=false</code> on the server host systems
	and follow the instructions below.
	</p>
	<p>JMeter uses Remote Method Invocation (RMI) as the remote communication mechanism. Therefore, you need
	to run the RMI Registry application (which is named, "<code>rmiregistry</code>") that comes with the JDK and is located in the "<code>bin</code>"
	directory. Before running <code>rmiregistry</code>, make sure that the following jars are in your system classpath:
	<ul>
	<li><code>JMETER_HOME/lib/ext/ApacheJMeter_core.jar</code></li>
	<li><code>JMETER_HOME/lib/jorphan.jar</code></li>
	<li><code>JMETER_HOME/lib/logkit-2.0.jar</code></li>
	</ul>
	The
	rmiregistry application needs access to certain JMeter classes. Run <code>rmiregistry</code> with no parameters. By default the
	application listens to port <code>1099</code>.</p>

	<p><b>Step 1b: Start the JMeter Server</b></p>
	<p>Once the RMI Registry application is running, start the JMeter Server.
	Use the "<code>-s</code>" option with the jmeter startup script ("<code>jmeter -s</code>").</p>

	<p>Steps 2 and 3 remain the same.</p>
	</subsection>
	<subsection name="&sect-num;.2 Tips" anchor="tips">
	<p>
	JMeter/RMI requires a connection from the client to the server. This will use the port you chose, default <code>1099</code>.<br/>
	JMeter/RMI also requires a reverse connection in order to return sample results from the server to the client.<br/>
	This will use a high-numbered port. <br/>
	This port can be controlled by jmeter property called <code>client.rmi.localport</code> in <code>jmeter.properties</code>.<br/>
	If there are any firewalls or other network filters between JMeter client and server,
	you will need to make sure that they are set up to allow the connections through.
	If necessary, use monitoring software to show what traffic is being generated.
	</p>
	<p>If you're running Suse Linux, these tips may help. The default installation may enable the firewall. In that case,
	remote testing will not work properly. The following tips were contributed by Sergey Ten.</p>
	<p>If you see connections refused, turn on debugging by passing the following options.</p>
	<source>
	rmiregistry -J-Dsun.rmi.log.debug=true \
	-J-Dsun.rmi.server.exceptionTrace=true \
	-J-Dsun.rmi.loader.logLevel=verbose \
	-J-Dsun.rmi.dgc.logLevel=verbose \
	-J-Dsun.rmi.transport.logLevel=verbose \
	-J-Dsun.rmi.transport.tcp.logLevel=verbose \
	</source>
	<p>Since JMeter 2.3.1, the RMI registry is started by the server; however the options can still be passed in from the JMeter command line.
	For example: "<code>jmeter -s -Dsun.rmi.loader.logLevel=verbose</code>" (i.e. omit the <code>-J</code> prefixes).
	Alternatively the properties can be defined in the <code>system.properties</code> file.
	</p>
	<p>The solution to the problem is to remove the loopbacks <code>127.0.0.1</code> and <code>127.0.0.2</code> from <code>/etc/hosts</code>.
	What happens is <code>jmeter-server</code> can't connect to rmiregistry if <code>127.0.0.2</code> loopback is not available.
	Use the following settings to fix the problem.</p>
	<p> Replace</p>
	<source>`dirname $0`/jmeter -s "$@"</source>
	<p> With</p>
	<source>
	HOST="-Djava.rmi.server.hostname=[computer_name][computer_domain] \
	-Djava.security.policy=`dirname $0`/[policy_file]" \
	`dirname $0`/jmeter $HOST -s "$@"
	</source>
	<p>Also create a policy file and add <code>[computer_name][computer_domain]</code> line to <code>/etc/hosts</code>.</p>

	<p>In order to better support SSH-tunneling of the RMI communication channels used
	in remote testing, since JMeter 2.6:</p>
	<ul>
	<li>a new property "<code>client.rmi.localport</code>" can be set to control the RMI port used by the RemoteSampleListenerImpl</li>
	<li>To support tunneling RMI traffic over an SSH tunnel as the remote endpoint using a port on the local machine,
	loopback interface is now allowed to be used if it has been specified directly using the Java System Property
	"<code>java.rmi.server.hostname</code>" parameter.</li>
	</ul>
	</subsection>
	<subsection name="&sect-num;.3 Using a different port" anchor="portchange">
	<p>By default, JMeter uses the standard RMI port <code>1099</code>. It is possible to change this. For this to work successfully,
	all the following need to agree:</p>
	<ul>
	<li>On the server, start <code>rmiregistry</code> using the new port number</li>
	<li>On the server, start JMeter with the property <code>server_port</code> defined</li>
	<li>On the client, update the <code>remote_hosts</code> property to include the new remote <code>host:port</code> settings</li>
	</ul>

	<p>Since Jmeter 2.1.1, the jmeter-server scripts provide support for changing the port.
	For example, assume you want to use port <code>1664</code> (perhaps <code>1099</code> is already used).</p>
	On Windows (in a DOS box)
	<source>
	C:\JMETER> SET SERVER_PORT=1664
	C:\JMETER> JMETER-SERVER [other options]
	</source>
	On Unix:
	<source>
	$ SERVER_PORT=1664 jmeter-server [other options]
	</source>
	[N.B. use upper case for the environment variable]
	<p>
	In both cases, the script starts rmiregistry on the specified port,
	and then starts JMeter in server mode, having defined the "<code>server_port</code>" property.
	</p>
	<p>
	The chosen port will be logged in the server <code>jmeter.log</code> file (<code>rmiregistry</code> does not create a log file).
	</p>
	</subsection>

	<subsection name="&sect-num;.4 Using a different sample sender" anchor="sendermode">
	<p>
	Listeners in the test plan send their results back to the client JMeter which writes the results to the specified files
	By default, samples are sent back synchronously as they are generated.
	This can affect the maximum throughput of the server test; the sample result has to be sent back before the thread can
	continue.
	There are some JMeter properties that can be set to alter this behaviour.
	</p>
	<dl>
	<dt><code>mode</code></dt><dd>sample sending mode - default is <code>StrippedBatch</code> since 2.9. This should be set on the client node.
	<dl>
	<dt><code>Standard</code></dt><dd>send samples synchronously as soon as they are generated</dd>
	<dt><code>Hold</code></dt><dd>hold samples in an array until the end of a run. This may use a lot of memory on the server and is discouraged.</dd>
	<dt><code>DiskStore</code></dt><dd>store samples in a disk file (under <code>java.io.temp</code>) until the end of a run.
	The serialised data file is deleted on JVM exit.</dd>
	<dt><code>StrippedDiskStore</code></dt><dd>remove responseData from successful samples, and use DiskStore sender to send them.</dd>
	<dt><code>Batch</code></dt><dd>send saved samples when either the count (<code>num_sample_threshold</code>) or time (<code>time_threshold</code>) exceeds a threshold,
	at which point the samples are sent synchronously.
	The thresholds can be configured on the server using the following properties:
	<dl>
	<dt><code>num_sample_threshold</code></dt><dd>number of samples to accumulate, default <code>100</code></dd>
	<dt><code>time_threshold</code></dt><dd>time threshold, default 60000 ms = 60 seconds</dd>
	</dl>

	See also the Asynch mode, described below.</dd>
	<dt><code>Statistical</code></dt><dd>send a summary sample when either the count or time exceeds a threshold.
	The samples are summarised by thread group name and sample label.
	The following fields are accumulated:
	<ul>
	<li><code>elapsed time</code></li>
	<li><code>latency</code></li>
	<li><code>bytes</code></li>
	<li><code>sample count</code></li>
	<li><code>error count</code></li>
	</ul>
	Other fields that vary between samples are lost.
	</dd>
	<dt><code>Stripped</code></dt><dd>remove responseData from successful samples</dd>
	<dt><code>StrippedBatch</code></dt><dd>remove responseData from successful samples, and use Batch sender to send them.</dd>
	<dt><code>Asynch</code></dt><dd>samples are temporarily stored in a local queue. A separate worker thread sends the samples.
	This allows the test thread to continue without waiting for the result to be sent back to the client.
	However, if samples are being created faster than they can be sent, the queue will eventually fill up,
	and the sampler thread will block until some samples can be drained from the queue.
	This mode is useful for smoothing out peaks in sample generation.
	The queue size can be adjusted by setting the JMeter property
	<code>asynch.batch.queue.size</code> (default <code>100</code>) on the server node.
	</dd>
	<dt><code>StrippedAsynch</code></dt><dd>remove responseData from successful samples, and use Async sender to send them.</dd>
	<dt><code>Custom implementation</code></dt><dd>set the mode parameter to your custom sample sender class name.
	This must implement the interface <code>SampleSender</code> and have a constructor which takes a single
	parameter of type <code>RemoteSampleListener</code>.
	</dd>
	</dl>
	</dd>
	</dl>
	<note><code>Stripped</code> mode family strips <code>responseData</code> so this means that some Elements that rely
	on the previous <code>responseData</code> being available will not work.<br/>
	This is not really a problem as there is always a more efficient way to implement this feature.
	</note>
	<p>The following properties apply to the <code>Batch</code> and <code>Statistical</code> modes:</p>
	<dl>
	<dt><code>num_sample_threshold</code></dt><dd>number of samples in a batch (default <code>100</code>)</dd>
	<dt><code>time_threshold</code></dt><dd>number of milliseconds to wait (default 60 seconds)</dd>
	</dl>
	</subsection>


	<subsection name="&sect-num;.5 Dealing with nodes that failed starting" anchor="retries">
	<p>
	For large-scale tests there is a chance that some part of remote servers will be unavailable or down.
	For example, when you use automation script to allocate many cloud machines and use them as generators,
	some of requested machines might fail booting because of cloud's issues.
	Since JMeter 2.13 there are new properties to control this behaviour.
	</p>
	<p>
	First what you might want is to retry initialization attempts in hope that failed nodes just slightly delayed their boot.
	To enable retries, you should set <code>client.tries</code> property to total number of connection attempts.
	By default it does only one attempt. To control retry delay, set the <code>client.retries_delay</code> property
	to number of milliseconds to sleep between attempts.
	</p>
	<p>
	Finally, you might still want to run the test with those generators that succeeded initialization and skipping failed nodes.
	To enable that, set the <code>client.continue_on_fail=true</code> property.
	</p>
	</subsection>

	</section>

	</body>
	</document>