【作者】 庞东贺；

【导师】 杨孝宗；

【作者基本信息】 哈尔滨工业大学 ， 计算机科学与技术， 2011， 硕士

【摘要】 随着云计算、数据中心、虚拟化逐渐深入并影响着信息化进程,高端容错计算机扮演着越来越重要的角色,高端容错计算机系统对可用性有很高的要求,因此对评测高端容错计算机可用性虽然任务艰巨但是具有重要意义。故障注入方法又是评测系统有效性的重要的实验方法,所以需要对故障注入技术进行深入的了解和研究。本文对国内外一些比较经典的故障注入技术进行了研究对比分析,发现虽然已经这些工具都已经很成熟并得到广泛地应用,但是这些工具测试手段较为单一,不能对目标系统进行全面的测试,而且传统工具的目标系统普遍是IA-32位体系结构的。针对上述问题,本文提出基于组件技术集成各类故障注入工具,实现了对目标系统不同抽象层次上的故障注入的多样化,同时支持故障注入工具“即插即用”的功能;基于平台集成的故障注入工具设计开发相关性测试套件、冗余度测试套件、MTTR测试套件、可用度测试套件,以得到相关指标来有效地评测系统的可用性。手动生成测试用例延长了测试时间,同时平台的测试环境需要覆盖各种各样的平台,增加了环境搭建的复杂性。考虑上述手动测试存在的问题,本文基于IBM提供的开源的自动化测试框架STAF,探索建立分布式自动化平台,通过动态部署测试环境、测试执行自动化,减少人机交互的次数,提高自动化测试程度,降低人力物力的消耗。最后,本文给出MTTR测试套件和基于MCA故障注入工具的应用实例,并通过对比故障注入手动测试和自动化测试,验证了开发平台的有效性。更多还原

【Abstract】 As cloud computing, data center and virtualization arebecoming more advanced and affecting informationization, high-end fault-tolerant computers are playing a more and more important role. High-end fault-tolerant computer systems require a high degree of usability. Therefore although evaluating usability isdifficult, it is extremely useful. The faultinjection method is an important methodfor evaluating system effectiveness, so it needs to be studied and understood more deeply.This paper researches and analyzes domestic and foreign classic fault injection technology. It finds that although these tools are already very advanced and widely used, there are relatively few means of testing. These meanscan’t thoroughly test the target systems, which are commonly IA - 32-bit architectures. In view of the above problems, this paper develops a platform which integrates all kinds of faults injection tools on the target system, which diversifies the means of testing on different levels. At the same time it supports the "plug and play" capability of fault injected tools.Based on the fault injected tools integrated into the platform, designing and developing the correlation test suite, the redundancy test suite, the MTTR test suite, the availability test suite gets the related indicators to effectively evaluate the system availability.Generating test cases manually extends the time of test. The testing environment of the fault injection platform needs to cover a wide range of platforms, which also increases the complexity of building environment. Considering the manual testing problems, this paper establishes a distributed and automation platform based on Software Testing Automation Framework (STAF). Through the testing environment dynamic deployment, test task automation control and the reduction of the number of the human-computer interaction, improve the degree of testing automation and reduce the consumption of manpower and material resources. Finally, this paper gives the application of the MTTR test suite and the fault injected tool based on MCA.Tthe effectiveness of developed platform has been verified by using the implemented fault-injection platform and test automation system to do the manual and automation tests.更多还原