海量下的分布式文件系统测试平台设计与实现

Design and realization of test platform for distributed file system under the massive storage

摘要:

为了对海量存储中的分布式文件系统的容错能力进行测试,设计并实现了基于海量存储下的自动化分布式文件系统测试平台,采用从服务层和目录文件层进行故障注入的技术,其中包括服务器失效和目录文件操作失败等多种故障. 该平台集成了注入故障工具、系统监测工具和工作负载工具,工作人员可通过平台的主控端进行评测. 通过实验数据表明,该测试平台能够有效评测不同文件系统的容错性能,并对不同存储系统的各项技术指标进行比较.

Abstract:

To evaluate distributed fault-tolerance, an automated test platform is designed and realized based on distributed fault-tolerance for massive storage, which contains a system monitor, a workload injection module and two fault-injection tools based on server level and directory and file operation. This platform can inject different levels faults including server level, directory level and files level, and can finish the test work by the master node. Experimental results show that the platform can compare performance index of the target system with different fault tolerance.