A Disjoint-Partitioning Approach to Enhancing Metamorphic Testing of DBMS

Abstract

Owing to big data, DBMS testing faces the oracle problem, that is, it is difficult to verify execution results against expected outcomes. Rigger and Su applied metamorphic testing to alleviate the challenge. We propose a disjoint-partitioning approach to extend their work. We have conducted an empirical case study on OceanBase, the DBMS associated with the world's fastest online transaction processing system. Even though Ocean- Base has been extensively tested and widely used in the industry, we have unveiled various hidden failures and crashes.