A debugging approach for Big Data applications in Pharo

Big Data applications are more and more popular; they typically analyze big sets of data from different domains. Many frameworks exist for programmers to develop and execute their Big Data applications such as Hadoop Map/Reduce and Apache Spark. However, very few debugging support is currently provided in those frameworks. When an error hap- pens, developers are lost in trying to understand what happened from the information provided in log files. Alternatively, few solutions allow to replay the execution, but they are slow and time-consuming. In this paper, we present an online approach to debug Big Data applications. We first introduce Port, a framework on top of Hadoop Yarn that al- lows to deploy and execute Pharo Map/Reduce applications. We debug applications deployed on such framework using IDRA, a novel online debugger for Pharo applications. With IDRA the running application can be debugged in a centralized way, and the code of the application can be dynamically updated to fix bugs.