<p>With the development of society, big data technology has been integrated into people’s production and life. The reliability of software (or components) for big data systems has received attention from people. So far, there has been relatively little research on reliability modeling and assessment of the software for big data systems. Considering the notable rise in the quantity of errors found in the later phases of testing of open source big data system software during the software development and test, this paper proposes a software reliability model based on a Weibull–Weibull distribution. The experimental results verified the effectiveness of the proposed model and the accuracy of predicting remaining faults in the software. The model proposed in this study can help developers to assess software reliability during the development of open source big data systems.</p>