HPjmeter

Change the command line of your Java application to use -Xeprof.

To collect Xeprof data during the entire execution of the launched Java application, use:

$ java -Xeprof <yourApp>

You can send the Xeprof output to a specified file using the file= keyword as follows:

$ java -Xeprof:file=<yourApp>.eprof <yourApp>

The <pid> will be inserted automatically in the file name, for example, <yourApp><pid>.eprof.

To collect Xeprof data for a specified time interval, there are two options.

	NOTE: If you are running JDK 1.5.0.04 or later, the command-line option is not required in order to capture eprof data. Instead you can toggle eprof data gathering on and off by sending signals to the currently running Java VM. One log file is produced per sample period; the name for the log file is java<pid>_<startTime>.eprof. The SIGUSR2 signal toggles the recording of eprof data. Use the following process to gather eprof data for specific periods: Send SIGUSR2 to the Java VM process. The Java VM will begin recording eprof data. Send SIGUSR2 to the Java VM process. The Java VM will flush eprof data and close the log file. See Profiling with Zero Preparation in the HPjmeter User's Guide for more information.

Run the application to create a data file.

Start the console from a local installation on your client machine. For example, here are two different ways:

$ $JAVA_HOME/bin/java <heap_settings> -jar /opt/hpjmeter/lib/HPjmeter.jar

$ /opt/hpjmeter/bin/hpjmeter

Click File—>Open File to browse for and open the data file.

A profile analysis screen opens, displaying a set of tabs containing summary and graphical metric data. The following screen shows an example:

Figure 1-5 HPjmeter — Profile Data

Click among the tabs to view available metrics. Use the Metrics or Estimate menus to select additional metrics to view. Each metric you select opens in a new tab. Hover your mouse over each category in the cascading menu to reveal the relevant metrics for that category. The following screen shows the available metrics for the threads/locks category:

Figure 1-6 HPjmeter — Threads/Locks Metrics

When running on HP-UX, add the -Xverbosegc option to the Java command line. (On non-HP-UX platforms, use the -Xloggc option instead.)

	NOTE: On JDK 5.014 (or later) and JDK 6.0.02 (or later), the command-line option is not required to capture Xverbosegc data. Instead, you can toggle Xverbosegc on and off by sending a SIGPROF signal to JVM. For more information, see “Allocation Site Statistics and Zero Preparation Xverbosegc”.

Run the application to create a data file.

Start the console from a local installation on your client machine. For example, here are two different ways:

$ $JAVA_HOME/bin/java <heap_settings> -jar /opt/hpjmeter/lib/HPjmeter.jar

$ /opt/hpjmeter/bin/hpjmeter

Click File—>Open File to browse for and open the data file.

A GC viewer screen opens and displays a set of tabs containing metric data. The following figure shows the garbage collection analysis screen:

Figure 1-7 HPjmeter — Garbage Collection Analysis

Set the SHLIB_PATH environment variable to include the location of the HPjmeter agent library as appropriate for 32 or 64-bit Java VM.

The following examples show how to set this variable in both the csh and the ksh for the different libraries.

To select the PA-RISC 32-bit library:

(csh) $ setenv SHLIB_PATH /opt/hpjmeter/lib/PA_RISC2.0
(ksh) $ export SHLIB_PATH=/opt/hpjmeter/lib/PA_RISC2.0       

To select the PA-RISC 64-bit library:

(csh) $ setenv SHLIB_PATH /opt/hpjmeter/lib/PA_RISC2.0W
(ksh) $ export SHLIB_PATH=/opt/hpjmeter/lib/PA_RISC2.0W      

To select the Integrity 32-bit library:

(csh) $ setenv SHLIB_PATH /opt/hpjmeter/lib/IA64N
(ksh) $ export SHLIB_PATH=/opt/hpjmeter/lib/IA64N       

To select the Integrity 64-bit library:

(csh) $ setenv SHLIB_PATH /opt/hpjmeter/lib/IA64W
(ksh) $ export SHLIB_PATH=/opt/hpjmeter/lib/IA64W

Confirm that the node agent is running. With a standard installation, the node agent should be running as a daemon on the system where it was installed. A node agent must be running before the console can connect to a managed node to discover applications and open monitoring sessions.

To verify that the node agent is running, use the following ps command:

$ ps -ef | grep node 

 $ JMETER_HOME/bin/nodeagent -daemon 

If the node agent is not running, follow these steps to enable it:

$ /opt/hpjmeter/bin/nodeagent

Start the Java application with the Java VM agent. For example, to start the myapp application on JDK 1.5, enter (For JDK 6.0, replace “java1.5” with “java6”):

$ /opt/java1.5/bin/java -Xms256m -Xmx512m -agentlib:jmeter myapp   

$ /opt/java1.4/bin/java -Xms256m -Xmx512m  \   
-Xbootclasspath/a:$JMETER_HOME/lib/agent.jar -Xrunjmeter myapp       

Start the HPjmeter console by entering the command:

$ /opt/hpjmeter/bin/hpjmeter

Select Connect from the File menu or select the Connect to Server icon []. The following screen is displayed:

Figure 1-8 HPjmeter — Connecting to Server

In the Connect to Server dialog box, enter the host name where the Java application and corresponding node agent are running.

If the node agent was started on a nonstandard port, specify the port number in the Optional Port box.

Click Connect. The running Java VM for each application should appear in the console main screen pane marked with the symbol.

	NOTE: If there is a connection failure, the symbol is not displayed. Instead the symbol is displayed next to the server name to indicate the server connection failure. If this happens, verify the node agent is running on the specified server.

If you want to connect to several node agents, repeat the previous steps.

Double-click the Java VM icon in the data pane for the application that you want to monitor. This opens the Session Preferences dialog box shown in the following screen:

Figure 1-9 HPjmeter — Setting Session Preferences

Check the default settings for metrics, filters, and alerts, and enable the settings you want to activate.

Click OK. The Session Preferences screen closes and the newly Open Session is visible, marked by the icon. Refer to the following screen for an example:

Figure 1-10 HPjmeter — Collecting Metrics

Wait for the console to collect metrics. The length of time depends on the application size, the load imposed on the application, and the selected preferences. Typically, the wait will be from 5 to 30 minutes. Longer collection time gives you greater accuracy in the results.

Click the open session or time slice to highlight the data to be viewed.

Use the Monitor menu on the console main screen to select the desired metrics. Refer to the following screen for an example:

Figure 1-11 HPjmeter — Choosing Metrics to Monitor

Select a metric. A metric visualizer displaying the chosen data opens. Refer to the HPjmeter User’s Guide for details on individual metrics and how to interpret the data.

Start the console.

Start the node agent if it is not running as a daemon.

Start the sample application from the command line:

$ cd $JMETER_HOME/demo 
$ export LD_LIBRARY_PATH=$JMETER_HOME/lib 
$ java -agentlib:jmeter agent.jar -jar ML1.jar     

$ cd $JMETER_HOME/demo 
$ ../bin/run_simple_jvmagent -jar sample_program
   

In the console main screen, click Connect and enter the host name of the machine running the sample application. If you specified a port number when starting the node agent, use the same port number. Otherwise, leave the port number box empty.

An icon representing the host appears in the main screen. After a few moments, the console also shows the sample application as a child node of the host.

Double-click the application node to open a monitoring session with the application.

Click OK to accept the default settings for metrics, filters, and alerts.