网络存储设备RAID管理系统的设计与实现
|
摘要
近年来以网络技术和计算机为基础的信息产业获得了巨大的发展,许多新的应用,如视频采集服务、科学计算技术、数据存储中心等的出现,导致信息的存储方式产生了革命性的变化,网络存储技术成为当前主流的存储技术。
本课题来自于北京普天的系统研发部实施的网络存储设备项目其中的RAID(Redundant Array of Independent Disks)管理系统,该系统的作用是通过管理RAID卡和底层的硬盘,为上层的应用提供接口,屏蔽底层的硬件细节。RAID管理系统的主要功能是负责磁盘阵列的管理,并且实现逻辑空间的管理。在本系统中需要构建可扩展的体系结构来实现对不同RAID卡的兼容,实现多卡集成,并为用户提供命令行方式的操作界面。
系统是在Linux环境下开发的,开发语言是C和C++。本论文首先阐述了系统的设计思想:实现在多卡的环境下进行磁盘阵列的管理。在多卡集成方面,使用了层次化设计的思想,结合适配器模式统一了不同卡的接口,并且使用消息类的方式简化了接口的设计。磁盘阵列管理中的各个功能如下,ARRAY管理:ARRAY简单来说是一组物理盘,通过把若干物理盘构建为一个ARRAY,可以方便构建逻辑驱动器;逻辑驱动器管理:逻辑驱动器构建在ARRAY之上,通过指定RAID级别,卷大小等信息构建更高效安全的逻辑盘;热备盘管理:为逻辑驱动器指定热备,热备是在物理盘坏了的情况下,可以自动顶替坏盘,保证系统继续工作,所以热备的使用可有效的提高系统可靠性;物理盘管理:查看物理盘的状态信息;配置信息与告警信息管理:逻辑驱动器的配置信息的导入和导出,以及告警信息的查询。系统使用的是命令行的操作方式。最后论文介绍了系统的类详细设计和各个类的实现方法。
论文的最后介绍了测试工具和方法。测试证明,RAID管理系统满足功能需求,可扩展性高,维护难度和成本低,在整个存储设备中可以稳定可靠的工作,使设备的性能和安全性提高。
In recent years, the information industry based on computer and web technology has an unprecedented development, and brings many new applications such as video services, scientific computing and data centers, which have led to a revolutionary change in the area of data storage technology. And currently the network storage technology becomes the mainstream technology.
The RAID (Redundant Array of Independent Disks) management system is part of the network storage device implemented by the system application department of Beijing Potevio, and its fuction is managing the underlying RAID cards and hard disks in order to prowide applications interfaces for the upper level. By the means it shields the details of the underlying hardware. RAID management system is mainly responsible for the management of disk array, and the realization of the logic disk management. The system needs to build a scalable architecture to achieve the aim that the different RAID cards can be compatibility multi-card integration, and to provide users with command line interface.
The system was developed in the Linux environment, and the development language is C and C plus plus. At the beginning, the dissertation describes the aim of the system: the realization of disk array management in multi-cards environment. The system adopts design pattern method to design multi-card integration. The use of a hierarchical design idea and adapter pattern can unify the different interface, and the use of message classes could simplify the designing the interface. Disk array management functions are as follows, ARRAY management: ARRAY is simply a group of physical disk. Through the construction of a number of physical disks as a ARRAY, a logical drive can be easily built; logical drive management: building a logical drive on the ARRAY through defining RAID level, volume size and other information can enhance efficient and security; hot spare disk management: first designate hot spare for the logical drive, when physical disk fail it will automatically replace the bad disk and continue to work. This feather could effectively improve system reliability; physical disk management: the principal objective of it is to view the status of physical disk information; configuration information and alert information management: manage the import and export of logical drive configuration information, besides inquiries of alarm information. The system provides a command line interface. Then this dissertation introduces the design and implementation of the classes.
The end of dissertation describes the test tools and methods. Has been tested, the RAID management system could meet the functional requirements, and has feathers of scalability, low difficulty and cost of maintenance. Working in the storage devices the system is stable and reliable, and it successfully enhances performance and security of equipment.
本课题来自于北京普天的系统研发部实施的网络存储设备项目其中的RAID(Redundant Array of Independent Disks)管理系统,该系统的作用是通过管理RAID卡和底层的硬盘,为上层的应用提供接口,屏蔽底层的硬件细节。RAID管理系统的主要功能是负责磁盘阵列的管理,并且实现逻辑空间的管理。在本系统中需要构建可扩展的体系结构来实现对不同RAID卡的兼容,实现多卡集成,并为用户提供命令行方式的操作界面。
系统是在Linux环境下开发的,开发语言是C和C++。本论文首先阐述了系统的设计思想:实现在多卡的环境下进行磁盘阵列的管理。在多卡集成方面,使用了层次化设计的思想,结合适配器模式统一了不同卡的接口,并且使用消息类的方式简化了接口的设计。磁盘阵列管理中的各个功能如下,ARRAY管理:ARRAY简单来说是一组物理盘,通过把若干物理盘构建为一个ARRAY,可以方便构建逻辑驱动器;逻辑驱动器管理:逻辑驱动器构建在ARRAY之上,通过指定RAID级别,卷大小等信息构建更高效安全的逻辑盘;热备盘管理:为逻辑驱动器指定热备,热备是在物理盘坏了的情况下,可以自动顶替坏盘,保证系统继续工作,所以热备的使用可有效的提高系统可靠性;物理盘管理:查看物理盘的状态信息;配置信息与告警信息管理:逻辑驱动器的配置信息的导入和导出,以及告警信息的查询。系统使用的是命令行的操作方式。最后论文介绍了系统的类详细设计和各个类的实现方法。
论文的最后介绍了测试工具和方法。测试证明,RAID管理系统满足功能需求,可扩展性高,维护难度和成本低,在整个存储设备中可以稳定可靠的工作,使设备的性能和安全性提高。
In recent years, the information industry based on computer and web technology has an unprecedented development, and brings many new applications such as video services, scientific computing and data centers, which have led to a revolutionary change in the area of data storage technology. And currently the network storage technology becomes the mainstream technology.
The RAID (Redundant Array of Independent Disks) management system is part of the network storage device implemented by the system application department of Beijing Potevio, and its fuction is managing the underlying RAID cards and hard disks in order to prowide applications interfaces for the upper level. By the means it shields the details of the underlying hardware. RAID management system is mainly responsible for the management of disk array, and the realization of the logic disk management. The system needs to build a scalable architecture to achieve the aim that the different RAID cards can be compatibility multi-card integration, and to provide users with command line interface.
The system was developed in the Linux environment, and the development language is C and C plus plus. At the beginning, the dissertation describes the aim of the system: the realization of disk array management in multi-cards environment. The system adopts design pattern method to design multi-card integration. The use of a hierarchical design idea and adapter pattern can unify the different interface, and the use of message classes could simplify the designing the interface. Disk array management functions are as follows, ARRAY management: ARRAY is simply a group of physical disk. Through the construction of a number of physical disks as a ARRAY, a logical drive can be easily built; logical drive management: building a logical drive on the ARRAY through defining RAID level, volume size and other information can enhance efficient and security; hot spare disk management: first designate hot spare for the logical drive, when physical disk fail it will automatically replace the bad disk and continue to work. This feather could effectively improve system reliability; physical disk management: the principal objective of it is to view the status of physical disk information; configuration information and alert information management: manage the import and export of logical drive configuration information, besides inquiries of alarm information. The system provides a command line interface. Then this dissertation introduces the design and implementation of the classes.
The end of dissertation describes the test tools and methods. Has been tested, the RAID management system could meet the functional requirements, and has feathers of scalability, low difficulty and cost of maintenance. Working in the storage devices the system is stable and reliable, and it successfully enhances performance and security of equipment.
引文
1张江陵,冯丹.海量信息存储.北京科学出版社, 2005:45～54
2 G.A.Gilbson, R.V.Meter. Network Attached Storage Architecture. Communications of the ACM, 2006, 43(11):37～45
3 J.R.Heath, P.J.Yakutis. High speed storage area networks using a fibre channel arbitrated loop interconnect. IEEE Network, 2007, 14(2):51～56
4宋宝莉.网络存储系统的设计与实现.电脑与信息技术, 2003, 21(6):31~33
5 G.A.Gibson, J.Wilkes. Strategic Directions in Computing Research Working Group on SCSI in Large-Scale Computing, Self_mamanging Network-attached Storage. ACM Computing, 1996, 28(4):30～43
6 Marc Farley. SAN存储区域网络(第2版).孙功星等译.北京机械工业出版社, 2001:179～182
7 Tom Clark, Thomas Clark. Designing Storage Area Networks(1st edition). Addison-Wesley Professional, 1999:10～15
8 Lu Yinping. Performance study of RAID-based storage subsystems. Communications Magazine, IEEE, 2003, 41(8):76～82
9蔡皖东.基于SAN的高可用性网络存储解决方案.小型微型计算机系统, 2001, 22(3):284～287
10 Mario Apicella. SAN and NAS. InfoWorld, 2005, 24(1):35～39
11 Kalman Z.Meth, Julian Satran.Features of the iSCSI protocol.Communicatio- Ns Magazine, IEEE, 2007, 41(8):40～42
12孙社文.网络存储主流技术分析.北京工业职业技术学院报, 2005, (04):32~33
13陈嵩,郑华.深入分析RAID的应用.福建电脑, 2002, (12):27~30
14 Francisco Matias Cuenca-Acuna, Richard Martin. Autonomous replication for high availability in storage systems. In : Proceedings.22nd International Symposium on Reliable Distributed Systems, 2003:99～108
15胡劲草. RAID存储设备的发展及其应用.电子技术, 2006, (11):12~14
16 G F Hughes, J F Murray.Reliability and Security of RAIDStorage Systems and D2D Archives Using SATA Disk Drivers. ACMTransactions on Storage, 2004:20~34
17 G.A.Gibson, D.A.Patterson. Designing Disk Arrays for High Data Reliability. Journal of Parallel and Distributed Computing, 1993, 17(1):4～27
18 Jiang Guo-song, Xie Chang-sheng, Wang Bo. Research and Implementation of RAID Algorithm base on RAID Controller, Networking, Architecture, and Storage. 2007 (6):251~252
19冯丹,张江陵.构造高可靠性磁盘阵列结构的研究.计算机工程与科学, 1996, 18(3):43～47, 337～356
20 Cheng Peng, Zhang Jiangling. HUST_RAID:High Performance RAID in Real-Time System.IEEE Pacific Rim Conference on Communications,1999:59～62
21 G.A.Gibson, J.Wilkes. Strategic Directions in Computing Research Working Group on Storage I/O Issue in Large-Scale Computing, Self_mamanging Network-attachedStorage.ACM Computing, 1996, 28(4):30～43
22 Thomas Clark. Designing Storage system (1st edition).Addison-Wesley Professional, 1998:10～15
23 G.A.Gilbson, R.V.Meter. Network Attached Storage Architecture. Communications of the ACM, 2007, 43(11):37～45
24 J.R.Heath. Design File System Server on NAS. IEEE Network, 2004, 12(2):21～30
25冯丹,张江陵.多磁盘阵列构成海量存储系统的研究.计算机研究与发展, 2004, 37(2):263～264
26 Mario Apicella. FS management on NAS.InfoWorld, 2005, 24(1):35～39
27 Ning Wang.SCSI-based storage subsystems. Communications Magazine, IEEE, 2004, 41(8):76～82
28 Kalman Z.Meth, Julian Satran.Features of the NAS.Communications Magazine, IEEE, 2003, 41(8):40～42
29韩东平.国内存储技术发展现状报告.福建电脑报, 2008, (1):12~13
30陈平仲.硬件实现RAID与软件实现RAID的比较.现代计算机,2005,(1):10~13
31王沛,韩耀伟,刘斌,潘金贵. Linux中SoftwareRAID驱动程序的机制分析.小型微型计算机系统, 2001, 22(3):305～308
32刘晓光.基于网络的软磁盘阵列的可用性研究.南开大学图书馆, 1999:40~50
33赵德平,史桂颖.基于Linux网络块设备和软RAID技术的网络镜像.计算机工程,2007(18):23~25
34曹阳. RAID技术实现及发展.电脑学习, 2006,(04):20~22
35李东.磁盘阵列的安全管理.电力信息化, 2004,(6):18~19
36饶国林. Linux下磁盘阵列的实现及其高可用性的研究.华中科技大学, 2006:20~30
37李善平,陈文智. Linux内核指导.浙江大学出版社, 2002:95～97
38魏永明,耿岳,钟书毅. Linux设备驱动程序.中国电力出版社, 2006:21～44, 474～481
39康华,张波,陈莉君. Linux内核设计与实现.北京机械工业出版社, 2005:173～184
40马琦,郭玉东. Linux下网络块设备的设计与实现.微机发展, 2003, 13(5):12～13
41 Claudia Salzberg Rodriguez,Gordon Fischer,Steven Smolski. The Linux Kernel Primer : A Top-Down Approach for x86 and PowerPC Architectures. Prentice Hall PTR, 2006(7):150~170
42鲍庆,郭玉东,曹剑青. Linux2.4内核下网络通信的接口分析.信息工程大学学报, 2003, 4(2):29～32
43熊伟,王刚,刘璟.基于Linux系统的容双故障软磁盘阵列盘平台Deraid.计算机工程, 2005, 31(8):7～9
2 G.A.Gilbson, R.V.Meter. Network Attached Storage Architecture. Communications of the ACM, 2006, 43(11):37～45
3 J.R.Heath, P.J.Yakutis. High speed storage area networks using a fibre channel arbitrated loop interconnect. IEEE Network, 2007, 14(2):51～56
4宋宝莉.网络存储系统的设计与实现.电脑与信息技术, 2003, 21(6):31~33
5 G.A.Gibson, J.Wilkes. Strategic Directions in Computing Research Working Group on SCSI in Large-Scale Computing, Self_mamanging Network-attached Storage. ACM Computing, 1996, 28(4):30～43
6 Marc Farley. SAN存储区域网络(第2版).孙功星等译.北京机械工业出版社, 2001:179～182
7 Tom Clark, Thomas Clark. Designing Storage Area Networks(1st edition). Addison-Wesley Professional, 1999:10～15
8 Lu Yinping. Performance study of RAID-based storage subsystems. Communications Magazine, IEEE, 2003, 41(8):76～82
9蔡皖东.基于SAN的高可用性网络存储解决方案.小型微型计算机系统, 2001, 22(3):284～287
10 Mario Apicella. SAN and NAS. InfoWorld, 2005, 24(1):35～39
11 Kalman Z.Meth, Julian Satran.Features of the iSCSI protocol.Communicatio- Ns Magazine, IEEE, 2007, 41(8):40～42
12孙社文.网络存储主流技术分析.北京工业职业技术学院报, 2005, (04):32~33
13陈嵩,郑华.深入分析RAID的应用.福建电脑, 2002, (12):27~30
14 Francisco Matias Cuenca-Acuna, Richard Martin. Autonomous replication for high availability in storage systems. In : Proceedings.22nd International Symposium on Reliable Distributed Systems, 2003:99～108
15胡劲草. RAID存储设备的发展及其应用.电子技术, 2006, (11):12~14
16 G F Hughes, J F Murray.Reliability and Security of RAIDStorage Systems and D2D Archives Using SATA Disk Drivers. ACMTransactions on Storage, 2004:20~34
17 G.A.Gibson, D.A.Patterson. Designing Disk Arrays for High Data Reliability. Journal of Parallel and Distributed Computing, 1993, 17(1):4～27
18 Jiang Guo-song, Xie Chang-sheng, Wang Bo. Research and Implementation of RAID Algorithm base on RAID Controller, Networking, Architecture, and Storage. 2007 (6):251~252
19冯丹,张江陵.构造高可靠性磁盘阵列结构的研究.计算机工程与科学, 1996, 18(3):43～47, 337～356
20 Cheng Peng, Zhang Jiangling. HUST_RAID:High Performance RAID in Real-Time System.IEEE Pacific Rim Conference on Communications,1999:59～62
21 G.A.Gibson, J.Wilkes. Strategic Directions in Computing Research Working Group on Storage I/O Issue in Large-Scale Computing, Self_mamanging Network-attachedStorage.ACM Computing, 1996, 28(4):30～43
22 Thomas Clark. Designing Storage system (1st edition).Addison-Wesley Professional, 1998:10～15
23 G.A.Gilbson, R.V.Meter. Network Attached Storage Architecture. Communications of the ACM, 2007, 43(11):37～45
24 J.R.Heath. Design File System Server on NAS. IEEE Network, 2004, 12(2):21～30
25冯丹,张江陵.多磁盘阵列构成海量存储系统的研究.计算机研究与发展, 2004, 37(2):263～264
26 Mario Apicella. FS management on NAS.InfoWorld, 2005, 24(1):35～39
27 Ning Wang.SCSI-based storage subsystems. Communications Magazine, IEEE, 2004, 41(8):76～82
28 Kalman Z.Meth, Julian Satran.Features of the NAS.Communications Magazine, IEEE, 2003, 41(8):40～42
29韩东平.国内存储技术发展现状报告.福建电脑报, 2008, (1):12~13
30陈平仲.硬件实现RAID与软件实现RAID的比较.现代计算机,2005,(1):10~13
31王沛,韩耀伟,刘斌,潘金贵. Linux中SoftwareRAID驱动程序的机制分析.小型微型计算机系统, 2001, 22(3):305～308
32刘晓光.基于网络的软磁盘阵列的可用性研究.南开大学图书馆, 1999:40~50
33赵德平,史桂颖.基于Linux网络块设备和软RAID技术的网络镜像.计算机工程,2007(18):23~25
34曹阳. RAID技术实现及发展.电脑学习, 2006,(04):20~22
35李东.磁盘阵列的安全管理.电力信息化, 2004,(6):18~19
36饶国林. Linux下磁盘阵列的实现及其高可用性的研究.华中科技大学, 2006:20~30
37李善平,陈文智. Linux内核指导.浙江大学出版社, 2002:95～97
38魏永明,耿岳,钟书毅. Linux设备驱动程序.中国电力出版社, 2006:21～44, 474～481
39康华,张波,陈莉君. Linux内核设计与实现.北京机械工业出版社, 2005:173～184
40马琦,郭玉东. Linux下网络块设备的设计与实现.微机发展, 2003, 13(5):12～13
41 Claudia Salzberg Rodriguez,Gordon Fischer,Steven Smolski. The Linux Kernel Primer : A Top-Down Approach for x86 and PowerPC Architectures. Prentice Hall PTR, 2006(7):150~170
42鲍庆,郭玉东,曹剑青. Linux2.4内核下网络通信的接口分析.信息工程大学学报, 2003, 4(2):29～32
43熊伟,王刚,刘璟.基于Linux系统的容双故障软磁盘阵列盘平台Deraid.计算机工程, 2005, 31(8):7～9