Data Intensive (DI) applications are becoming more and more important in several fields of science, economy, and even in our normal life. Unfortunately, even if some technological frameworks are available for their development, we still lack solid software engineering approaches to support their development and, in particular, to ensure that they offer the required properties in terms of availability, throughput, data loss, etc.. In this paper we report our action research experience in developing-testing-reengineering a specific DI application, Hegira4Cloud, that migrates data between widely used NoSQL databases. We highlight the issues we have faced during our experience and we show how cumbersome, expensive and time-consuming the developing-testing-reengineering approach can be in this specific case. Also, we analyse the state of the art in the light of our experience and identify weaknesses and open challenges that could generate new research in the areas of software design and verification.