To facilitate this exercise, we built a simple android application which would trigger
OutOfMemoryError. We named this app ‘BuggyApp.’ Do you like the name? This app has just one page, which contains only one button: “Add Objects.” When you click on this button, the app would start to add very large string objects to an array list in an infinite while loop. When large string objects are infinitely added to an array list, it will result in
Figure 1: Android App that simulates OutOfMemoryError
Figure 2: Android App experiencing OutOfMemoryError
Few seconds after you click on “Add Objects” button, application will crash with
OutOfMemoryError, as shown below:
Figure 3: Android OutOfMemoryError crash stack trace
How to diagnose OutOfMemoryError?
Now let’s get to the interesting part — how to diagnose
OutOfMemoryError? It’s just two easy simple steps, my friend:
- Capture Android Heap Dumps
- Analyze Android Heap Dumps
1. Capture Android Heap Dumps
First step is to capture heap dumps from the android app. Heap Dump is a snapshot of memory, which contains information about the objects in the memory, their references, their size… Here is an article which summarizes 3 different options to capture heap dumps from the android app. You can use any one of the options that is convenient for you to capture heap dumps. We p used option #2 mentioned in the article to capture the heap dump from this ‘BuggyApp.’
2. Analyze Android Heap Dumps
Second step is to analyze the captured heap dump. To analyze android heap dumps, we used the free online tool: HeapHero. This tool analyzes android heap dumps and points out potential memory leak suspects. Besides identifying memory leaks, HeapHero also identifies the amount of memory wasted due to poor programming practices and recommends solutions to fix the same. Since it’s an online tool, you don’t have to do any downloading, installation, setup…. All you need to do is to upload the heap dump file that was captured in step #1 to HeapHero.
Now we uploaded the heap dump file captured in step #1 to HeapHero. HeapHero generated this beautiful report. In the report, there is a section: Large Objects. This section reports all the large objects that are residing in the memory.
Figure 4: Large Objects sitting in memory reported by HeapHero
From the report, you could see the
android.view.inputmethod.InputMethodManager.sInstance object to occupy 96.7% of overall memory. This is a significant size for any object. In fact, HeapHero has in-built intelligence to identify potential memory leaking objects. Based on this intelligence, HeapHero also marked
android.view.inputmethod.InputMethodManager.sInstance object as potential memory leak suspect.
When we clicked on the
android.view.inputmethod.InputMethodManager.sInstance hyperlink in the report, it started to show stack trace/code path of the leaking objects.
Figure 5: Code Path of memory leaking objects
You could see the
buggycompany.com.buggyapp.MainActivity object holding
java.util.ArrayList object, which in turn is holding on to large amount of string objects (96.6%). Ah ha, exact lines of code that we wrote in ‘BuggyApp’ to trigger
That’s it my friend. With these two simple steps, you might be able to solve complex
OutOfMemoryError and memory leak problems. We hope this article will help you to isolate the exact lines of code that are triggering memory problems in your app.
Author bio: Ram Lakshmanan
Every single day, millions & millions of people in North America — bank, travel, and commerce — use the applications that Ram Lakshmanan has architected. Ram is an acclaimed speaker in major conferences on scalability, availability, and performance topics. Recently, he has founded a startup, which specializes in troubleshooting performance problems.