Weblogic Authors: Yeshim Deniz, Elizabeth White, Michael Meiner, Michael Bushong, Avi Rosenthal

Related Topics: IBM Cloud

IBM Cloud: Article

Hunting Java Memory Leaks

Failure to manage memory can significantly hinder performance

In a production environment, memory leaks can force organizations to add more memory and hardware resources. They can even cause an application to crash unexpectedly. In theory, Java memory leaks should not emerge as a development or production issue because the garbage collector is responsible for memory management.

Built-in Java memory management enables developers to sidestep tedious memory management considerations and is a fundamental reason why Java is a relatively straightforward development environment. Practically, however, the garbage collector function is unable to accurately address all memory management considerations, due to programming oversights and Java language program loopholes that allow memory leaks.

The garbage collector's job is to periodically collect unreferenced objects. Java memory leaks typically occur because the garbage collector reserves, or allocates, memory for each object associated with a program - whether or not the program uses the object. The garbage collector, as a result, reserves memory for each object, but does not have the intelligence to reallocate memory reserved for an idle object to another command. Consequently, the garbage collector falsely assumes that memory is being used by objects. The objects that are not being used but still hold references are not collected, thus leading to memory leaks.

Failure to adequately manage memory can significantly hinder the performance of your Java and J2EE applications. This column will outline the root causes of memory leaks and detail best practices for identifying and eliminating memory leaks in development and production to ensure robust Java and J2EE application performance.

Four Leading Causes of Memory Leaks

Memory management is a complex process, and the source of a memory leak is not always evident. Outlined below are four leading causes of memory leaks.

Static collection classes, such as HashMap and Vector, are susceptible to memory leaks because they are designed to hold other referenced objects. Static key words (as shown in Listing 1) are likely to cause memory leaks because static variables remain in memory as long as the application runs, regardless of its object creation and destruction.

There are several listener types available in J2EE that provide notification when an event occurs. Until a J2EE application receives event notification, any storage held by that application will not be garbage-collected by the Java Virtual Machine (JVM). Web application listener events are new in the Servlet 2.3 Specifi-ca-tion. These listeners include servlet creation and invalidation as well as application startup and shutdown. The two listener classes, javax.Servlet.servletContextListener and javax.servlet.http.HttpSessionListener, when not implemented correctly, can lead to long-running applications that hold storage. To avoid this problem, keep the storage associated with servlet sessions small to minimize the effect on the heap. The Graphics Interchange Format (GIF), for example, should not be held in a session object that is kept in session using a listener.

Physical connections, such as database and network connections, are not going to be garbage-collected automatically unless they are disconnected explicitly. Java database connections are typically initialized using the DataSource.getConnection() method, which is used to handle physical connections. These physical connections must be freed when they are no longer required by the application using the close() method. Because the scope of such connections is independent of the JVM, the garbage collector does not have the ability to affect a physical connection. When your application server uses the connection pool, the DataSource.getConnection() method allows you to release connections back to the pool so other transactions can use the physical connection to access the database.

Connection pools bring advantages to application development by reducing the overhead of an application by physically connecting to and disconnecting from a database. The physical connections are maintained by the application server and managed from a pool that is farmed out to each application that retrieves data from a database. Using a connection pool limits the number of physical database connections. Therefore, it is very important for an application to release its connection once it has finished sharing information with a database.

Similar to the physical connections mentioned above, the Java Native Interface (JNI) allows Java code to run with other languages, such as C and C++. The references created in the C or C++ application environments should be unreferenced explicitly. Because the C or C++ storage is not allocated from the JVM Heap, it will not be garbage-collected by the JVM. Because the JVM does not have the ability to de-allocate JNI references, JNI memory must be managed directly by the C or C++ application.

A Best-Practice Approach to Hunting Memory Leaks

It is not always easy to detect memory leaks by scanning your code. Not all of the programmers involved with a J2EE project may have the time or experience to code memory management appropriately. Although many project managers face this issue, there is no standard solution to this problem because nearly every Java environment is unique. There are, however, guidelines that development teams should follow to optimize memory performance. An overview of best practices for ensuring a high level of memory performance throughout each phase of the "build, run, manage" application infrastructure life cycle follows.

Listing 2 illustrates how programmers can overlook memory leaks. The helper object in Listing 2 is a memory leak until another class removes it from the LongReferent class. Listing 2 shows that the garbage collector is not always responsible for collecting memory. The memory leak in this program can be avoided by removing the helper object from LongReferent explicitly, by the same class or by some other class. Objects stored in a HashMap are removed using the clear() or remove(object key) HashMap class method calls. Once the reference is removed from the HashMap, it is a candidate for garbage collection. Always double-check that the references are released, especially when you write static collection classes, listeners, database access code, and JNI code.

One approach to avoiding memory leaks is to write both "create method" and "remove method" commands in each class, then check for remove-method usage. In the unit-testing phase, you can search for create-method and remove-method usage in the code. The memory code is functioning properly if there is an equal number of calls for the create method and remove method commands. If the number of calls for these methods don't match, you should fix the code by adding or removing methods wherever they are missing.

To achieve an added level of control over memory and garbage collection from the program, J2EE introduced the java.lang.ref package. For example, the WeakHashMap class java.lang.ref package uses weak keys to reference objects within the garbage collection. An object is available for the garbage collector when the WeakHashMap key is invalid. The garbage collector can make a key invalid if the memory is low, allowing the object it references to be freed. The garbage collector moves the key through three stages: finalizable, finalized, and reclaimed. It is not unlikely that the application could try to retrieve an object only to determine the object has been reclaimed by a garbage-collection cycle. This process works effectively for objects that can be re-created and easily collected whenever memory is running low.

Listing 3 shows the usage of the WeakHashMap class. Listing 3 is similar to Listing 2, except that it uses the WeakHashMap instead of the HashMap class. The difference between the two classes is that HashMap contains strongly referenced keys, which cannot be removed until the program unreferences them explicitly. The WeakHashMap class contains weakly referenced keys that can be removed by the garbage collector without unreferencing, or removing, keys in the program when memory is running low. The garbage collector algorithm determines the removal of weakly referenced keys. This varies by JVM.

The garbage collector generally removes unused, weakly referenced objects first so that it can allocate the unused memory for new objects. The program also checks for null keys and refreshes them with keys and values when it uses WeakHashMap. As part of this process, the garbage collector works in the background to remove the keys automatically. These classes are important to consider, especially when you implement memory-hog functionality - such as caching modules - that enables cached objects to be easily re-created. This approach can be used to replace the HashMap calls in the first example.

Even when following best practices, it is virtually impossible to avoid memory leaks in medium or large projects in which many programmers are writing thousands of lines of code. Fortunately, there are effective performance tools available to pinpoint memory leaks automatically. Some products are built into development solutions, and others are stand-alone tools. Some of the newest development tools include profilers, which enable developers to monitor performance issues from the same development integrated development environment. Profiler solutions from Candle Corp. and other vendors can pinpoint memory leaks automatically as part of a broader development tool. An effective profiler tool must enable developers to:

  1. Identify which method allocated the memory
  2. Clarify if the memory was released by the garbage collector and how long it was allocated
  3. Determine which objects are holding the reference to the memory; Listing 1 shows that the HashMap is holding the memory reference
The bar chart in Figure 1 shows the top 10 allocated object references. The table below tells the programmer what program allocated the memory and if it had been garbage-collected. An object reference tab that shows the method holding the reference to the memory is also available. Many of today's tools can make it very easy to pinpoint memory leaks and, therefore, avoid production outages.

Although profilers excel at identifying memory leaks and other performance problems, developers and testers must be able to use the tools effectively. Memory problems that are overlooked in development will likely cascade in the production phase. Profiles provide development teams with a significant amount of performance data, which, in turn, may create more development cycles in the project due to the additional work associated with finely tuning memory management. Added production cycles may not be feasible for projects that have tight deadlines.

Fortunately, monitoring tools in production enable users to track J2EE application performance in real-world environments. Production profiling solutions from Candle, IBM, and other vendors monitor resource utilization for memory consumption. These tools provide visibility into the J2EE application environment that enables users to proactively improve performance and avoid production downtime.


Memory leaks often destabilize an application in production by consuming large quantities of memory or causing the JVM to crash. Moreover, whenever the system is low on memory the garbage collector tries to collect the memory more frequently, thus diminishing application performance. During development, programmers must pay special attention to memory leak problems associated with static collection classes, listeners, physical connections, and JNIs. The latest profiling tools identify memory problems long before the testing and production phases.

Production monitoring tools enable you to track application health and view otherwise-hidden performance metrics. Organizations are better able to tune memory management, which can improve the overall production stability and uptime. Organizations that address memory leaks by using a combination of development best practices and profiling and management tools can optimize memory management, which can increase the performance of their J2EE applications.

More Stories By Lloyd Hagemo

Lloyd Hagemo is a senior director for Candle Corporation’s Application Infrastructure Management Group. He is responsible for WebSphere tools development. Lloyd has led the successful development of more than 20 products for the WebSphere environment, including operating system utilities, network performance and tuning products, WebSphere MQ configuration and management tools, and application integration solutions.

More Stories By Ravi Kalidindi

Ravi Kalidindi is a senior software engineer for Candle Corporation’s Application Infrastructure Management Group. Ravi has worked with Java and J2EE since its inception on a wide variety of J2EE projects and has published several articles that focus on Java and J2EE best practices

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.

IoT & Smart Cities Stories
DXWorldEXPO LLC announced today that ICOHOLDER named "Media Sponsor" of Miami Blockchain Event by FinTechEXPO. ICOHOLDER gives detailed information and help the community to invest in the trusty projects. Miami Blockchain Event by FinTechEXPO has opened its Call for Papers. The two-day event will present 20 top Blockchain experts. All speaking inquiries which covers the following information can be submitted by email to [email protected] Miami Blockchain Event by FinTechEXPOalso offers sp...
DXWordEXPO New York 2018, colocated with CloudEXPO New York 2018 will be held November 11-13, 2018, in New York City and will bring together Cloud Computing, FinTech and Blockchain, Digital Transformation, Big Data, Internet of Things, DevOps, AI, Machine Learning and WebRTC to one location.
@DevOpsSummit at Cloud Expo, taking place November 12-13 in New York City, NY, is co-located with 22nd international CloudEXPO | first international DXWorldEXPO and will feature technical sessions from a rock star conference faculty and the leading industry players in the world. The widespread success of cloud computing is driving the DevOps revolution in enterprise IT. Now as never before, development teams must communicate and collaborate in a dynamic, 24/7/365 environment. There is no time t...
Headquartered in Plainsboro, NJ, Synametrics Technologies has provided IT professionals and computer systems developers since 1997. Based on the success of their initial product offerings (WinSQL and DeltaCopy), the company continues to create and hone innovative products that help its customers get more from their computer applications, databases and infrastructure. To date, over one million users around the world have chosen Synametrics solutions to help power their accelerated business or per...
When talking IoT we often focus on the devices, the sensors, the hardware itself. The new smart appliances, the new smart or self-driving cars (which are amalgamations of many ‘things'). When we are looking at the world of IoT, we should take a step back, look at the big picture. What value are these devices providing. IoT is not about the devices, its about the data consumed and generated. The devices are tools, mechanisms, conduits. This paper discusses the considerations when dealing with the...
Charles Araujo is an industry analyst, internationally recognized authority on the Digital Enterprise and author of The Quantum Age of IT: Why Everything You Know About IT is About to Change. As Principal Analyst with Intellyx, he writes, speaks and advises organizations on how to navigate through this time of disruption. He is also the founder of The Institute for Digital Transformation and a sought after keynote speaker. He has been a regular contributor to both InformationWeek and CIO Insight...
Machine learning has taken residence at our cities' cores and now we can finally have "smart cities." Cities are a collection of buildings made to provide the structure and safety necessary for people to function, create and survive. Buildings are a pool of ever-changing performance data from large automated systems such as heating and cooling to the people that live and work within them. Through machine learning, buildings can optimize performance, reduce costs, and improve occupant comfort by ...
Digital Transformation is much more than a buzzword. The radical shift to digital mechanisms for almost every process is evident across all industries and verticals. This is often especially true in financial services, where the legacy environment is many times unable to keep up with the rapidly shifting demands of the consumer. The constant pressure to provide complete, omnichannel delivery of customer-facing solutions to meet both regulatory and customer demands is putting enormous pressure on...
Bill Schmarzo, Tech Chair of "Big Data | Analytics" of upcoming CloudEXPO | DXWorldEXPO New York (November 12-13, 2018, New York City) today announced the outline and schedule of the track. "The track has been designed in experience/degree order," said Schmarzo. "So, that folks who attend the entire track can leave the conference with some of the skills necessary to get their work done when they get back to their offices. It actually ties back to some work that I'm doing at the University of San...
CloudEXPO New York 2018, colocated with DXWorldEXPO New York 2018 will be held November 11-13, 2018, in New York City and will bring together Cloud Computing, FinTech and Blockchain, Digital Transformation, Big Data, Internet of Things, DevOps, AI, Machine Learning and WebRTC to one location.