Diagnosing memory leaks using graph mining on heap dumps

Memory leaks are caused by software programs that prevent the reclamation of memory that is no longer in use. They can cause significant slowdowns, exhaustion of available storage space and, eventually, application crashes. Detecting mem-ory leaks is challenging because real-world applications are built on multiple layers of software frameworks, making it difficult for a developer to know whether observed references to objects are legitimate or the cause of a leak. We present a graph mining solution to this problem wherein we ana-lyze heap dumps to automatically identify subgraphs which could represent potential memory leak sources. Although heap dumps are commonly analyzed in existing heap pro-filing tools, our work is the first to apply a graph grammar mining solution to this problem. Unlike classical graph min-ing work, we show that it suffices to mine the dominator tree of the heap dump, which is significantly smaller than the underlying graph. Our approach identifies not just leak-ing candidates and their structure, but also provides aggre-gate information about the access path to the leaks. We demonstrate several synthetic as well as real-world exam-ples of heap dumps for which our approach provides more insight into the problem than state-of-the-art tools such as Eclipse’s MAT