存储数据保护Raid技术DDP说明.pptx

上传人:b****1 文档编号:1425167 上传时间:2022-10-22 格式:PPTX 页数:20 大小:2.45MB
下载 相关 举报
存储数据保护Raid技术DDP说明.pptx_第1页
第1页 / 共20页
存储数据保护Raid技术DDP说明.pptx_第2页
第2页 / 共20页
存储数据保护Raid技术DDP说明.pptx_第3页
第3页 / 共20页
存储数据保护Raid技术DDP说明.pptx_第4页
第4页 / 共20页
存储数据保护Raid技术DDP说明.pptx_第5页
第5页 / 共20页
点击查看更多>>
下载资源
资源描述

存储数据保护Raid技术DDP说明.pptx

《存储数据保护Raid技术DDP说明.pptx》由会员分享,可在线阅读,更多相关《存储数据保护Raid技术DDP说明.pptx(20页珍藏版)》请在冰豆网上搜索。

存储数据保护Raid技术DDP说明.pptx

,存储高级数据保护技术-DDP说明,1,Volumes,SANtricityRAIDProtection,VolumegroupsRAID0,1,10,5,6IntermixRAIDlevelsVariousgroupsizesDynamicdiskpoolsMin11SSDsMax120SSDsUpto10diskpoolspersystem,2,NetAppConfidential,VolumeGroups,Volumes,HostLUNs,SSDs,DiskPool,SSDs,HostLUNs,SANtricityRAIDLevels,RAID0stripedRAID5datadisksandrotatingparity,RAID1(10)mirroredandstripedRAID6(P+Q)datadisksandrotatingdualparity,3,NetAppConfidential,Block-levelstripingwithadistributedparity,TraditionalRAIDVolumes,DiskdrivesorganizedintoRAIDgroupsVolumesresideacrossthedrivesinaRAIDgroupPerformanceisdictatedbythenumberofspindlesHotsparessitidleuntiladrivefailsSparecapacityis“stranded”,24-drivesystemwith210-drivegroups(8+2)and4hotspares,4,TraditionalRAIDDriveFailure,DataisreconstructedontohotspareSingledriveresponsibleforallwrites(bottleneck)Reconstructionhappenslinearly(onestripeatatime)Allvolumesinthatgrouparesignificantlyimpacted,24-drivesystemwith210-drivegroups(8+2)and4hotspares,5,TheProblem,TheLarge-Disk-DriveChallengeStaggeringamountsofdatatostore,protect,accessSomesiteshavethousandsoflarge-capacitydrivesDrivefailuresarecontinual,particularlywithNL-SASdrivesProductionI/OisimpactedduringrebuildsUpto40%inmanycasesAsdrivecapacitiescontinuetogrow,traditionalRAIDprotectionispushedtoitslimitDrivetransferrateshavenotkeptupwithcapacitiesLargerdrivesequallongerrebuildsanywherefrom10+hourstoseveraldays,6,4TB+,DynamicDiskPools,MaintainSLAsduringdrivefailureStayinthegreenPerformancedropisminimizedfollowingdrivefailureDynamicrebalancecompletesupto8xfasterthantraditionalRAIDinrandomenvironmentsandupto2xfasterinsequentialenvironmentsLargepoolofspindlesforeveryvolumereduceshotspotsEachvolumespreadacrossalldrivesinpoolDynamicdistribution/redistributionisanondisruptivebackgroundoperation,7,Balanced:

Algorithmrandomlyspreadsdataacrossalldrives,balancingworkloadandrebuildingifnecessary.,Easy:

NoRAIDoridlesparestomanageactivesparecapacityonalldrives.,Combiningeffort:

Alldrivesinthepoolsustaintheworkloadperfectforvirtualmixedworkloadsorfastreconstructionifneeded.,Flexible:

AddANY*numberofdrivesforadditionalcapacitysystemautomaticallyrebalancesdataforoptimalperformance.,TraditionalRAIDTechnology,InnovativeDynamicDiskPools,8,“WithDynamicDiskPools,youcanaddorlosediskdriveswithoutimpact,reconfiguration,orheadaches.”,*Aftertheminimumof11.,9,DataRebalancinginMinutesvs.Days,Hours,2.5Days,1.3Days,Typicalrebalancingimprovementsarebasedona24-diskmixedworkload,Morethan4Days,BusinessImpact,99%ExposureImprovement,MaintainbusinessSLAswithadrivefailure,RAIDLevelComparison,10,NetAppConfidential,DynamicDiskPoolsOverview,DDPdynamicallydistributesdata,sparecapacity,andparityinformationacrossapoolofSSDsAlldrivesareactive(noidlehotspares)SparecapacityisavailabletoallvolumesDataisdynamicallyrecreated/redistributedwheneverpoolsgrowsorshrinks,11,NetAppConfidential,DDP:

Simplicity,Performance,Protection,SimplifiedadministrationNoRAIDsetsorhotsparestomanageDataisautomaticallybalancedwithinpoolFlexiblediskpoolsizingoptimizescapacityutilizationConsistentperformanceDataisdistributedthroughputthepool(nohotspots)PerformancedropisminimizedduringadriverebuildSignificantlyfasterreturntooptimalstateRelentlessdataprotectionSignificantlyfasterrebuildtimesasdataisreconstructedthroughoutthediskpoolPrioritizedreconstructionminimizesexposure,12,NetAppConfidential,DDPInsight:

HowItWorks,EachDDPvolumeiscomposedofsomenumberof4GB“virtualstripes”calleddynamicstripesEachD-striperesidesonapseudo-randomlyselectedsetof10drivesfromwithinthepoolD-stripesareallocatedattimeofvolumecreationandallocatedsequentiallyonaper-volumebasis,13,NetAppConfidential,24-SSDpool,DDPSSDFailure,ForeachD-StripewhichhasdataonfailedSSD:

SegmentsonotherSSDsarereadtorecreatedataAnewSSDischosentowritesegmentsfromfailedSSDRebuildoperationsruninparallelacrossallSSDs,14,NetAppConfidential,DDPMultipleDiskFailure,IftwoSSDshavefailedsystemwillrebuildcriticalsegmentsfirstBrownandLightBlue,15,NetAppConfidential,IfadditionalSSDsfailnewcriticalsegmentswillbeidentifiedandrebuiltBlue,OrangeandPink,DDPAddingSSDsToPool,AddasingleSSDoraddmultipleSSDssimultaneouslyImmediatelyrebalancesdatatomaintainequilibriumSegmentsarejustmoved(doesnotreconstructRAID),16,NetAppConfidential,23-SSDpool,DynamicStripe:

ACloserLook,DataiswrittenwithinaD-stripeusingRAID6EachRAID6stripeis1MB(8+2with128Ksegmentsize)EachD-stripehas4,096traditionalRAID6stripesAD-pieceisonedrivesworthofD-stripedata,17,NetAp

展开阅读全文
相关资源
猜你喜欢
相关搜索

当前位置:首页 > 考试认证 > IT认证

copyright@ 2008-2022 冰豆网网站版权所有

经营许可证编号:鄂ICP备2022015515号-1