JMX (Java Management Extensions) Interview Preparation Guide

Download PDF

The application JVM is configured with the following options:
☛ com.sun.management.jmxremote
☛ com.sun.management.jmxremote.port=1088
☛ com.sun.management.jmxremote.authenticate=false
☛ com.sun.management.jmxremote.ssl=false

Here are a few advantages of Setter:
☛ Natural for Java Beans
☛ Avoids overly long constructors
☛ Can be inherited
☛ Easier to understand in that properties have names
☛ Constructor args don't have names

You have to expose the JMX port to the host and even used Sun specific options while running the Java process.
You are able to telnet to the host ip and the exposed JMX port, which tells that it is accessible. But I can't figure out a way to use JConsole and connect to the JMX service running in the container.
☛ Djava.rmi.server.hostname=$JMX_HOSTNAME
☛ Dcom.sun.management.jmxremote.port=$JMX_PORT
☛ Dcom.sun.management.jmxremote.rmi.port=$JMX_PORT

Here are 5 steps:
☛ Read bean file XML (or process annotations). Build internal "map" of beans and determine dependency tree.
☛ Post process the definitions - e.g. resolve ${variables}
☛ Allocate objects - dependency inject constructor args and invoke setter.
☛ Perform any initialisation - 3 approachs: @PostConstruct, init-method, InitializingBean
☛ Bean post-processors - e.g. wrap with proxy, @Required

A pointcut expression that has a name so it can be reused.
Resons To use:
☛ Annotation: @Pointcut applied to a dummy method - name of the dummy method is the name of the pointcut expression.
☛ XML aop:pointcut id=".." expression=".."

Types of advice:
☛ Before
☛ AfterReturning
☛ AfterThrowing
☛ After
☛ Around

he code that implements the cross cutting-concern that will be inserted at each join-point.

Joint-point is an expression that defines zero or more join-points.

A requirement that cuts across all the natural modules of your application.
Examples:
Tracing, security, transactions, business rules.
Two problems with cross-cutting concerns.
☛ Code tangling - method doing too many unrelated tasks
☛ Code scattering - code duplication leading to maintenance headache

Bean is not allocated until it is actually needed - dependency injected - or requested via getBean(). specify lazy="true" (false by default). Only useful if bean genuinely may never be used in most runs of the application.

Scopes the bean within the bean that uses it. One less bean id in the set of bean names.

Two examples:
☛ Nested or inner beans
☛ Configuration beans

5 scopes defined in Spring:
☛ Singleton
☛ Prototype
☛ Session - only valid in a Web application context
☛ Request - ditto
☛ Custom

It is created by Spring automatically, any setters are called on the factory then its getObject() method is called to allocate the actual bean.

Here are a few advantages:
☛Constructor
☛Mandatory
☛Mutable
☛Natural Java approach.

<constructor-arg… and <property…

There are nested elements within each bean definition.

Different collaborators for different environments can be injected - test, production, test-database, real database. This allows classes to be tested independent of its collaborators (makes it easy to pass in stub or mock objects).

A pattern (design) whereby collaborators are passed into an object from the outside - the object does not allocate them for itself.

Copy {WLHOME}/wlserver_**/server/lib/wlfullclient.jar file into {agent home}/opt/groovy-1.8.8/lib Then the Groovy script (that migrated from my JMX Java application) works.

When Tomcat is running as a service on Windows, you do not need setenv.bat, you need to use tomcat8w.exe and GUI toll for adding options.

No JMX-related output in the logs. I can connect to the JVM with jconsole using the supplied credentials, browse to the configured mbean, jackhammer the Refresh button and get new values. Splunk, however, gets nothing, and leaves no indication as to why. Searching for "host=redacted" or "source='jmx://redacted'" yields nothing.
This is not a complicated setup, nor is it a complicated app. Single server test environment, just trying to get JMX data into Splunk directly instead of having to do some nasty search-time field extraction after indexing data output from JMXTrans text files. It's currently running on JDK 1.7, I tried 1.6 for science to no avail. No idea how to proceed since I'm not getting any errors and there doesn't appear to be a way to raise the log level.

You have to install the jmx_ta on a 6.2 forwarder and getting the following error"
12-04-2014 16:08:02.884 +0000 ERROR ModularInputs - Introspecting scheme=jmx: script running failed (exited with code 1).
12-04-2014 16:08:02.884 +0000 ERROR ModularInputs - Unable to initialize modular input "jmx" defined inside the app "jmx_ta":
Introspecting scheme=jmx: script running failed (exited with code 1).

Install Splunk 6 on your server which contains about 6 different JMX instances. you want to hook all of them up into our Splunk JMX plugin.
It is as easy as editing the config.xml file and adding a new jmx server host? Currently this is:
<jmxserver host="[[system name]]" jvmDescription="common_raw" jmxport="53632">

A managed bean - sometimes simply referred to as an MBean - is a type of JavaBean, created with dependency injection. Managed Beans are particularly used in the Java Management Extensions technology. But, with the Java EE 6 specification provides for a more detailed meaning of a managed bean.