Complete Data Deletion Based on Hadoop Distributed File System

Abstract

There is a disk residual problem for the deletion mechanism of the Hadoop file storage system. This paper proposes a "plug and play" deletion method to achieve the purpose of completely deleting data. After simulation and recovery experiments, this method can effectively destroy the original data. At the same time, this method can be applied to some Big data clusters which are based on Hadoop. It does not need to redeploy the cluster and can be directly deployed in the already running big data cluster.