摘要
任何文件系统都具有拍摄快照的能力,快照使系统管理员在系统故障的情况下进行数据恢复。HDFS(Hadoop Distributed File System)是一个流行的大型集群商用服务器的分布式文件系统,该文为HDFS提出了一个低开销的快照解决方案。该解决方案,避免使用复杂的分布式快照算法,采用新的快照数据结构解决了在HDFS系统中的数据拷贝和分布问题,设计了一个基于选择性的追加拷贝快照解决方案。
The ability to take snapshots is an essential functionality of any file system, as snapshots enable system administrators to perform data recovery in case of failure. HDFS(Hadoop Distributed File System)is a popular distributed file system for large clusters of commodity servers. This paper presents a low overhead snapshot solutions for HDFS. The solution obviates the need for complex distributed snapshot algorithms,takes new snapshot data structures to data replication and distribution in HDFS. we have designed, a fast and efficient snapshot solution based on selective-copy-on-appends.
作者
杨彬
YANG Bin (Department of Information Technology, Liaoning Academy of Goverance, Shenyang 100161, Chiina)
出处
《电脑知识与技术》
2015年第2期74-78,共5页
Computer Knowledge and Technology
关键词
HDFS
快照
选择性的追加拷贝
低开销
Hadoop Distributed File System
snapshot
selective-copy-on-appends
low overhead