In this paper we propose a data mining technique for the discovery of intrusions in big data. To achieve our objective, we first reviewed the different data mining works and tools to our knowledge for the extraction of data from big data. Secondly, we chose a honeypot (honeyD) from a set (of honeypots) based on well-defined criteria. Thirdly, we combined this honeypot (honeyD) with different classification algorithms (decision trees and clustering such as k-means, DBSCAN to identify possible intrusions into the databases) in a functional architecture in which, we have presented and explained the role of each of its components. The implementation of our proposal shows that the combination of the honeypot with these different clustering algorithms gives convincing results which make it possible to detect possible intrusions in the data big databases.