Argumentation is a major topic in the study of Artificial Intelligence. Since the first edition in 2015, advancements in solving (abstract) argumentation frameworks are assessed in competition events, similar to other closely related problem solving technologies. In this paper, we report about the design and results of the Second International Competition on Computational Models of Argumentation, which has been jointly organized by TU Dresden (Germany), TU Wien (Austria), and the University of Genova (Italy), in affiliation with the 2017 International Workshop on Theory and Applications of Formal Argumentation. This second edition maintains some of the design choices made in the first event, e.g. the I/O formats, the basic reasoning problems, and the organization into tasks and tracks. At the same time, it introduces significant novelties, e.g. three additional prominent semantics, and an instance selection stage for classifying instances according to their empirical hardness.