基于备份的RAID6在线重构框架
|
摘要
针对国内外现有研究一直解决不了重负载持续访问下RAID6重构性能急剧恶化的问题,提出"利用外部存放的备份数据来加速磁盘阵列RAID6在线重构"的思想。构建基于备份的RAID6在线重构框架,实现该框架的原型系统,并进行相应测试。此框架利用备份系统所提供的恢复带宽将处于最近一次备份时间点的版本数据整合热备盘;利用磁盘阵列RAID6所提供的重构带宽将自最近一次备份时间点之后已修改数据重构至热备盘;完成磁盘阵列RAID6重构过程,使得磁盘阵列RAID6恢复正常运行。测试结果说明了该框架相对于现有重构方法显著改善了磁盘阵列RAID6的重构性能和服务性能。
Regarding present problems that the rapid deterioration of reconstruction performance under the continuous heavy workload has never been solved by both the domestics and foreign researches,this thesis proposed for the idea of"accelerating the online reconstruction of RAID6 by the backup data in external storage ". We built the online reconstruction framework of RAID6 based on backup,and completed the prototype system of this framework,and did the corresponding test. The framework utilized the restore bandwidth provided by the backup system to integrate the version data at the latest backup time point into the hot spare disk,and utilized the reconstruction bandwidth provided by RAID6 to reconstruct the modified data after the latest backup time point to the hot spare disk. It completed the reconstruction process of the disk array RAID6,and restored the disk array RAID6 to normal operation. The test result shows that compared with the existing reconstruction method,the framework greatly improves the reconstruction performance and service performance of RAID6.
Regarding present problems that the rapid deterioration of reconstruction performance under the continuous heavy workload has never been solved by both the domestics and foreign researches,this thesis proposed for the idea of"accelerating the online reconstruction of RAID6 by the backup data in external storage ". We built the online reconstruction framework of RAID6 based on backup,and completed the prototype system of this framework,and did the corresponding test. The framework utilized the restore bandwidth provided by the backup system to integrate the version data at the latest backup time point into the hot spare disk,and utilized the reconstruction bandwidth provided by RAID6 to reconstruct the modified data after the latest backup time point to the hot spare disk. It completed the reconstruction process of the disk array RAID6,and restored the disk array RAID6 to normal operation. The test result shows that compared with the existing reconstruction method,the framework greatly improves the reconstruction performance and service performance of RAID6.
引文
[1]Rhea S,Wells C,Eaton P,et al.Maintenance-free global data storage[J].IEEE Internet Computing,2001,5(5):40-49.
[2]Pinheiro E,Weber W D,Barroso L A.Failure Trends in a Large Disk Drive Population.[C]//Usenix Conference on File and Storage Technologies.San Jose,Usa.USENIX,2007:17-28.
[3]Reddy A L N,Chandy J,Banerjee P.Design and Evaluation of Gracefully Degradable Disk Arrays[J].Journal of Parallel&Distributed Computing,1993,17(1-2):28-40.
[4]Qin X,Miller E L,Schwarz S J T J E.Evaluation of distributed recovery in large-scale storage systems[C]//IEEE International Symposium on High PERFORMANCE Distributed Computing,2004.Proceedings.IEEE,2004:172-181.
[5]Wu X,Li J,Kameda H.Reliability Modeling of Declustered-Parity RAID Considering Uncorrectable Bit Errors[J].Ieice Transactions on Fundamentals of Electronics Communications&Computer Sciences,1997,80(8):1508-1514.
[6]Luo X,Shu J,Zhao Y.Shifted Element Arrangement in Mirror Disk Arrays for High Data Availability during Reconstruction[C]//International Conference on Parallel Processing.IEEE,2012:178-188.
[7]Wan J,Wang J,Xie C,et al.Formula Not Shown-RAID:Parallel RAID Architecture for Fast Data Recovery[J].IEEE Transactions on Parallel&Distributed Systems,2014,25(6):1638-1647.
[8]Xie P,Huang J Z,Cao Q,et al.V2-Code:A new non-MDSarray code with optimal reconstruction performance for RAID-6[C]//2013 IEEE International Conference on Cluster Computing(CLUSTER).IEEE,2013:1-8.
[9]Xie P,Huang J Z,Dai E W,et al.An efficient data layout scheme for better I/O balancing in RAID-6 storage systems[J].Frontiers of Information Technology&Electronic Engineering,2015,16(5):335-345.
[10]Fu Y,Shu J,Luo X,et al.Short Code:An Efficient RAID-6 MDS Code for Optimizing Degraded Reads and Partial Stripe Writes[J].IEEE Transactions on Computers,2016,66(1):127-137.
[11]Fu G,Thomasian A,Han C,et al.Rebuild Strategies for Redundant Disk Arrays.[C]//Symposium on Mass Storage Systems.2004:223-226.
[12]Tian L,Feng D,Jiang H,et al.PRO:A Popularity-based Multi-threaded Reconstruction Optimization for RAID-Structured Storage Systems.[C]//Usenix Conference on File and Storage Technologies,FAST 2007,February 13-16,2007,San Jose,Ca,Usa.DBLP,2007:277-290.
[13]陈金忠,姚念民,蔡绍滨.基于NAND闪存的高性能和可靠的PRAID-6[J].电子学报,2015,43(6):1211-1217.
[14]刘靖宇,谭毓安,薛静锋,等.S-RAID中基于连续数据特征的写优化策略[J].计算机学报,2014,37(3):721-734.
[15]Lei T,Hong J,Dan F,et al.Implementation and Evaluation of a Popularity-Based Reconstruction Optimization Algorithm in Availability-Oriented Disk Arrays[C]//MASS Storage Systems and Technologies,2007.MSST 2007.IEEE Conference on.IEEE,2007:233-238.
[16]Lumb C R,Schindler J,Ganger G R,et al.Towards higher disk head utilization:extracting free bandwidth from busy disk drives[C]//Conference on Symposium on Operating System Design&Implementation.USENIX Association,2000:7-7.
[17]Thereska E,Schindler J,Bucy J,et al.A framework for building unobtrusive disk maintenance applications[C]//Proceedings of the third USENIX Conference on File and Storage Technologies,2004,2:213-226.
[18]Liu F,Pan W,Xie T,et al.PDB:A Reliability-Driven Data Reconstruction Strategy Based on Popular Data Backup for RAID4 SSD Arrays[M]//Algorithms and Architectures for Parallel Processing.2013:87-100.
[19]Khasymski A,Rafique M M,Butt A R,et al.On the Use of GPUs in Realizing Cost-Effective Distributed RAID[C]//IEEE,International Symposium on Modeling,Analysis and Simulation of Computer and Telecommunication Systems.IEEE Computer Society,2012:469-478.
[20]高玲玲,许胤龙,王英子,等.基于RAID6编码的校验盘故障修复算法[J].计算机应用与软件,2014,31(6):248-251,302.
[2]Pinheiro E,Weber W D,Barroso L A.Failure Trends in a Large Disk Drive Population.[C]//Usenix Conference on File and Storage Technologies.San Jose,Usa.USENIX,2007:17-28.
[3]Reddy A L N,Chandy J,Banerjee P.Design and Evaluation of Gracefully Degradable Disk Arrays[J].Journal of Parallel&Distributed Computing,1993,17(1-2):28-40.
[4]Qin X,Miller E L,Schwarz S J T J E.Evaluation of distributed recovery in large-scale storage systems[C]//IEEE International Symposium on High PERFORMANCE Distributed Computing,2004.Proceedings.IEEE,2004:172-181.
[5]Wu X,Li J,Kameda H.Reliability Modeling of Declustered-Parity RAID Considering Uncorrectable Bit Errors[J].Ieice Transactions on Fundamentals of Electronics Communications&Computer Sciences,1997,80(8):1508-1514.
[6]Luo X,Shu J,Zhao Y.Shifted Element Arrangement in Mirror Disk Arrays for High Data Availability during Reconstruction[C]//International Conference on Parallel Processing.IEEE,2012:178-188.
[7]Wan J,Wang J,Xie C,et al.Formula Not Shown-RAID:Parallel RAID Architecture for Fast Data Recovery[J].IEEE Transactions on Parallel&Distributed Systems,2014,25(6):1638-1647.
[8]Xie P,Huang J Z,Cao Q,et al.V2-Code:A new non-MDSarray code with optimal reconstruction performance for RAID-6[C]//2013 IEEE International Conference on Cluster Computing(CLUSTER).IEEE,2013:1-8.
[9]Xie P,Huang J Z,Dai E W,et al.An efficient data layout scheme for better I/O balancing in RAID-6 storage systems[J].Frontiers of Information Technology&Electronic Engineering,2015,16(5):335-345.
[10]Fu Y,Shu J,Luo X,et al.Short Code:An Efficient RAID-6 MDS Code for Optimizing Degraded Reads and Partial Stripe Writes[J].IEEE Transactions on Computers,2016,66(1):127-137.
[11]Fu G,Thomasian A,Han C,et al.Rebuild Strategies for Redundant Disk Arrays.[C]//Symposium on Mass Storage Systems.2004:223-226.
[12]Tian L,Feng D,Jiang H,et al.PRO:A Popularity-based Multi-threaded Reconstruction Optimization for RAID-Structured Storage Systems.[C]//Usenix Conference on File and Storage Technologies,FAST 2007,February 13-16,2007,San Jose,Ca,Usa.DBLP,2007:277-290.
[13]陈金忠,姚念民,蔡绍滨.基于NAND闪存的高性能和可靠的PRAID-6[J].电子学报,2015,43(6):1211-1217.
[14]刘靖宇,谭毓安,薛静锋,等.S-RAID中基于连续数据特征的写优化策略[J].计算机学报,2014,37(3):721-734.
[15]Lei T,Hong J,Dan F,et al.Implementation and Evaluation of a Popularity-Based Reconstruction Optimization Algorithm in Availability-Oriented Disk Arrays[C]//MASS Storage Systems and Technologies,2007.MSST 2007.IEEE Conference on.IEEE,2007:233-238.
[16]Lumb C R,Schindler J,Ganger G R,et al.Towards higher disk head utilization:extracting free bandwidth from busy disk drives[C]//Conference on Symposium on Operating System Design&Implementation.USENIX Association,2000:7-7.
[17]Thereska E,Schindler J,Bucy J,et al.A framework for building unobtrusive disk maintenance applications[C]//Proceedings of the third USENIX Conference on File and Storage Technologies,2004,2:213-226.
[18]Liu F,Pan W,Xie T,et al.PDB:A Reliability-Driven Data Reconstruction Strategy Based on Popular Data Backup for RAID4 SSD Arrays[M]//Algorithms and Architectures for Parallel Processing.2013:87-100.
[19]Khasymski A,Rafique M M,Butt A R,et al.On the Use of GPUs in Realizing Cost-Effective Distributed RAID[C]//IEEE,International Symposium on Modeling,Analysis and Simulation of Computer and Telecommunication Systems.IEEE Computer Society,2012:469-478.
[20]高玲玲,许胤龙,王英子,等.基于RAID6编码的校验盘故障修复算法[J].计算机应用与软件,2014,31(6):248-251,302.