HPE 3PAR軟件功能介紹

1、HPE 3PAR軟件功能介紹HPE 3PAR Software Overview3PAR Software and Features 3PAR Software Stack and Licensing 3PAR Adaptive Data Reduction Technologies 3PAR Dynamic and Adaptive Optimization3PAR Virtual Copy and Full Copy3PAR Virtual Lock3PAR Virtual Domain3PAR Priority Optimization 3PAR File Serving Options

2、23PAR Architectural Differentiation Purpose built on native virtualization33PAR StoreServ3PAR Single System Software (part of the base system) Autonomic PoliciesVirtual CopyFile PersonaOnline ImportDynamic Optimization3PAR fine-grained OSAdaptive Data ReductionAutonomic PoliciesManagement Integratio

3、nSystem ReporterPersistency FeaturesFlash CacheRecovery Manager Central App Suite Smart SANAdaptive OptimizationVirtual Domains3PAR Multi System Software (optional)Remote CopyPeer PersistenceStorage FederationCluster Extension3PAR Data at Rest Encryption (optional)Data Encryption LTUData at Rest Dri

4、ves 3PAR Hardware3PAR CPUs & ASICsActive Mesh Fast RAIDMixed WorkloadData Hashing for Adaptive Data ReductionAll-inclusive 3PAR Software SuitesFrame-based dramatically simplifies licensing4Frame license bundled with Base Controllers including:OS SuiteAdaptive Data Reduction (ADR)File Persona (FP)Vir

5、tual Copy (VC)Dynamic Optimization (DO)Adaptive Optimization (AO)Priority Optimization (PO)Virtual DomainsVirtual LockOnline ImportRecovery Manager Central (RMC)Recovery Manager Central App SuitesAll legacy Recovery ManagersSmart SANData at Rest EncryptionOptional frame license (1 x per array) inclu

6、ding:Data Encryption (requires self-encrypting drives SED)All-inclusive Single-System SoftwareOptional frame license (1 x per array) including:Remote Copy (RC)Peer Motion (PM)Peer Persistence (PP)Cluster Extension (CLX)Storage FederationAll-inclusive Multi-System Software3PAR Thin and Adaptive Data

7、Reduction Technologies3PAR Adaptive Data Reduction Technologies OverviewBuilt-in3PAR Storage is built from the ground up to support Thin Provisioning, Deduplication, Compression and Data Packing reducing the performance limitations that plague bolt-on thin provisioning and dedupe solutions. In-bandS

8、equences of data are analyzed inline by the 3PAR ASIC in 16kB chunks. Chunks of zeroes are not written to the backend, repeating chunks of same data can be deduplicated, other chunks can be compressed. Reservation-less3PAR ThP draws fine-grained increments from a single free space reservoir without

9、pre-dedication of any kind. Other vendors ThP implementation require a separate, pre-dedicated pool for each data service level.Integrated API for automatic thin reclamation in Symantec File System, VMware vSphere, Oracle ASM, Windows Server 2012, 2016 and othersGuaranteed efficiencySave up to 75% s

10、torage capacity with Adaptive Data Reduction and Flash. 63PAR ASICIntel multi-core CPU3PAR Adaptive Data Reduction Overview TCO and space efficiency without compromise7Buy less storage capacity Start Thin with Thin ProvisioningGet Thin withThin Conversion (Zero Detect)Stay Thin with Thin Persistence

11、Reduce Tech Refresh CostsIntegrated space reclamation2TB1TB16TB8TBLinuxPresented 24TBConsumed3TB + BufferBuffer00000000000000000000000000000011010001101110101101000110111010BeforeAfterFastHPE 3PAR StoreServ Adaptive Data Reduction technologiesTechnologies work together for optimal results8When used

12、together, duplicate pages are removed first and unique pages are then compressedRead the 3PAR Adaptive Data Reduction whitepaper to make best use of the technologies Express Index tablesZero Detect and Deduplication by the 3PAR ASICCompression by the Intel CPUData in cacheResulting data written to S

13、SDExpress ScanUnique dataCompressed and packed write data70+ GBTimeT10 UNMAP vmkfstools -y(768KB 42MB Coarse)Standard Thin ProvisioningSlow, Post-processOverhead60+ GBRapid, Inline ASIC Zero Detect3PAR Scalable ADR Technologies40+ GBTimeT10 UNMAP vmkfstools -y(16KB granularity)20+ GBVMware vSphere S

14、pace Reclamation9Fine-grained Reclaim FastCoarse Reclaim Slow3PAR In-line ADR TechnologyTraditional Storage with Space ReclaimXXXX0000000000000000000000000000000025GB25GB25GB25GB10GB10GBDatastores25GB25GB25GB25GB10GB10GB20GB20GB20GB20GBXX00000000000000000000000000000000XX000020GB VMDKs post deletion

15、s can consume 40+ GB 20GB VMDKs post deletions consume only 20+ GB0003PAR Thin Persistence Oracle IntegrationOracle auto-extend allows customers to save on database capacity with Thin Provisioning Database Capacity can get stranded after subsequent writes and deletes3PAR Thin Persistence and Oracle

16、ASM Storage Reclamation Utility (ASRU) can reclaim 25% or more stranded capacityAfter Tablefile shrink/drop or Database dropAfter a new LUN is added to ASM Disk GroupOracle ASM Utility compacts files and writes zeroes to free space3PAR Thin Built-In ASIC-based, zero-detection eliminates free spaceFr

17、om a DBA perspective:Non-disruptive does not impact storage performance Increase DB Miles Per Gallon10ASM with ASRUTraditional ArrayUnused space remainsZeroes are written0000000000Disk Group TablespaceTables00000000Oracle3PAR with Thin PersistenceFiles compacted by ASRUZeroes removed Space reclaimed

18、Linux space reclamation Automatic if the mounted file system supports periodic (online) discard i.e. it is mounted with “discard” option. (e.g. Red Hat 6.x ext4, SLES 11 SP2 Btrfs)For example in /etc/fstab:/dev/vgappP01/lvappP01 /app ext4 defaults,discard 2 2orUse batch discard via command fstrim(8)

19、.This approach is implemented via FITRIM ioctlThe file system itself has to support this ioctlAt this point ext4, ext3, xfs, btrfs and ocfs2 support itFor example:# fstrim v /myfs11Allocation units and alignmentImportant things to consider when implementing ADRThe granularity of 3PAR deduplication i

20、s 16 KiB and therefore the dedup efficiency is best when the blocks written are aligned to this granularityFor hosts that use file systems with tunable allocation units consider setting the allocation unit to 16 KiB or a multiple of 16 KiBFor Microsoft Windows the NTFS default allocation unit size i

21、s 4kiB for volumes up to 16TB. MS Article ID: 140365 /kb/140365?wa=wsignin1.0 Set the value to16kB or higher when formatting the NTFS volume; find the procedure here:/en-us/kb/2272294 Linux hosts using Btrfs, EXT4 and XFS file systems require no tuning EXT3 file systems align to 4 KiB boundaries and

22、 cannot be tuned; for maximum dedup efficiency migrate to EXT4 For applications that have tunable block sizes consider to setting the block size to 16 KiB or a multiple of 16 KiB12For more info also see the 3PAR Adaptive Data Reduction whitepaper 3PAR Optimization and Tuning Dynamic OptimizationAdap

23、tive Optimization133PAR Dynamic and Adaptive Optimization Manual or Automatic Tiering 14Tier 0 Tier 1 Tier 2 3PAR Dynamic Optimization3PAR Adaptive Optimization - Region Sub-LUN Block Movements between CPGs/Tiers based on PoliciesComplete LUN Movement between CPGs/TiersCPG 1CPG 2CPG 3CPG ACPG BCPG C

24、Online Virtual Volume tuning with Dynamic OptimizationNon-disruptively convert VVs within or between CPGsAfter a change to the CPG layoutSet size etc. After the installation of new drives of the same typeFor RAID level migration CPG to CPG15Tuning Example with Dynamic OptimizationTune a Virtual Volu

25、me from a NL CPG to an FC CPG16Iometer before tuneTune startedTune finishedIometer after tuneOnline Virtual Volume conversion with Dynamic OptimizationPart of Dynamic OptimizationNon-disruptively convert VVs between fully and thinly provisioning Within or across CPGs Thin VV with Dedup and/or Compre

26、ssion enabledOriginal volume can either be Deleted Saved with a prefix or suffix Saved with a new nameNote: Volumes to be migrated must not be in a Remote Copy relationship173PAR Optimization Dynamic OptimizationAdaptive Optimization183PAR Dynamic and Adaptive Optimization Manual or Automatic Tierin

27、g 19Tier 0 Tier 1 Tier 2 3PAR Dynamic Optimization3PAR Adaptive Optimization - Region Sub-LUN Block Movements between CPGs/Tiers based on PoliciesComplete LUN Movement between CPGs/TiersCPG 1CPG 2CPG 3CPG ACPG BCPG C3PAR Adaptive Optimization Improve Storage Utilization 20Traditional deploymentSingl

28、e pool of same disk drive type, speed and capacity and RAID levelNumber and type of disks are dictate by the max IOPS + capacity requirementsDeployment with 3PAR Adaptive OptimizationAn AO Virtual Volume draws space from 2 or 3 different tiers Each tier can be built on different CPGs, disk types, RA

29、ID level and number of disks Required IOPSRequired CapacityIO distribution0%100%0%100%High-speed media poolSingle pool of high-speed media Medium-speed media poolLow-speed media poolWasted spaceRequired IOPSRequired Capacity0%100%0%100%IOdistributionAddressing I/O Density with a tiered 3PAR system21

30、The following SSD IO densities can be achieved with 3PAR based on practical field information:% of total SSD net capacity% of total IO1332.550566108020903599System Reporter graphic3PAR Adaptive OptimizationImprove Storage Utilization 22Used Space GiBThis is an single tier arrayMost of the capacity h

31、as very low IO activity suitable for NL For some IO activity SSDs would be beneficialAdding a lower tier would decrease cost without compromising overall performance One tier without Adaptive Optimization Low activityHigh activityUsed Space GiBThree tiers with Adaptive Optimization runningTwo tiers

32、have been added and Adaptive Optimization enabledAdaptive Optimization has moved the least used Chunklets to the lower tier (red) and the heaviest used to the higher tier (blue)Low activityHigh activityNL drivesFC drivesSSD drivesFC drivesAccess per GiB per minAccess per GiB per min3PAR Virtual Copy

33、 and Full Copy233PAR Virtual Copy Snapshot at its best24Base VolumeSmartIndividually erasable and promotable Scheduled creation/deletionConsistency groups Copy on Write (CoW) for full and thinly provisioned VVRedirect on Write (RoW) for deduped VV on FlashThinNo reservation, non-duplicativeVariable

34、QoSReadyInstantaneously readable and/or writeableSnapshots of snapshots of Virtual Lock for retention of read-only snapsCan be created from SSMC, CLI, WEB API, SMI-S etc.Automated erase optionIntegratedHPE Recovery Manager Central App suites Backup Apps from HPE, Symantec, VEEAM, CommVault Up to 128

35、000 Virtual Volumes and Snapshots per array # of Snapshots Model 144357440c1044810400100237400c9642104009044104008910 10400859774008092720077771080076531040076487200761210400727910400723610400714110400709710400708010800704174006899108006761740066967200cTop Arrays WW as of February 2016Virtual Copies

36、100s of Snaps per Base Volume but only one CoW requiredAlso see the Virtual Copy Whitepaper3PAR Virtual Copy creation Copy on Write (CoW) for fully or thinly provisioned Virtual Volumes (FPVV and TPVV)25CPGsVirtual Volumes12345BaseRO SnapshotRW SnapshotCreate SnapshotRO SnapshotRW Snapshot3PAR Virtu

37、al Copy writesCopy on Write (CoW) for fully or thinly provisioned Virtual Volumes (FPVV and TPVV)26CPGsVirtual Volumes12345BaseRO SnapshotRW SnapshotWrite to Base VVCopy block to Snapshot CPGRedirect SnapshotsUpdate block Write to RW SnapshotCopy block to Snapshot CPGRedirect SnapshotsWrite16Write73

38、PAR Virtual Copy for Backup use caseOne week based on hourly snaps and an average daily change rate of 10%27Base Volume of 2 TB 24 Copies200 GB48 Copies200 GB72 Copies200 GB96 Copies200 GB120Copies200 GB144Copies200 GB168 Copies200 GBResults in 168 Virtual Copies and only 1.4 TB Snapshot Space neede

39、dMondayTuesdayWednesdayThursdayFridaySaturdaySundayGo back to SW and Features3PAR Full Copy V1 restorable copyPart of the base 3PAR OS 28Full physical point-in-time copy Provisionable after copy endsIndependent of base volumes RAID and physical layout propertiesFast resynchronization capabilityThin

40、Provisioning-awareFull copies can consume same physical capacity as thinly provisioned base volumeBase VolumeFull CopyFull CopyIntermediate Snapshot3PAR Full Copy V2 instantly accessible copy Part of the base 3PAR OS 29Share data quickly and easilyFull physical point-in-time copy Immediately provisi

41、onable to hostsIndependent of base volumes RAID and physical layout properties No resynchronization capabilityThin Provisioning-awareFull copies can consume same physical capacity as thinly provisioned base volumeBase VolumeFull CopyFull CopyIntermediate Snapshot3PAR Virtual Lock303PAR Virtual Lock3

42、PAR Virtual Lock Software prevents deletion of selected Virtual Volumes for a specified period of timeLocked Virtual Volumes cannot be deleted, even by a 3PAR Storage System administrator with the highest level of privileges. Note: mounted servers can still read, write and delete Files and Folders i

43、n the Volumes. Locked RO Virtual Copies cannot be deleted and overwritten (i.e. for compliance reasons) Because its tamper-proof, its also a way to avoid administrative mistakes.Supported withFat and Thin Virtual Volumes Full Copy, Virtual Copy and Remote Copy31Also see the Virtual Lock overview3PAR

44、 Virtual Lock32Easily set during snapshot creation by defining Retention and/or Expiration Time in a Volume Policy NoteLocked Virtual Volumes cannot be deleted, even by a 3PAR Storage System user with the highest level privilegesThe Virtual Lock license is required3PAR Virtual Domains33 What are 3PA

45、R Virtual Domains?34Multi-Tenancy with Traditional StorageMulti-Tenancy with 3PAR DomainsSeparate, Physically-Secured StorageShared, logically-secured 3PAR StorageAdmin AApp ADept. ACustomer AAdmin BApp BDept. BCustomer BAdmin CApp CDept. CCustomer CDomain AAdmin AApp ADept. ACustomer AAdmin BApp BD

46、ept. BCustomer BDomain BAdmin CApp CDept. CCustomer CDomain CRead the Virtual Domains Whitepaper: /v2/GetDocument.aspx?docname=4AA6-5240ENW 3PAR Virtual Domains for Multi-Tenancy Security35VirtualDomain nVirtualDomain AHostsAccessVolumesCPGParametersQoSHostsAccessVolumesCPGParametersQoSProvides fine

47、-grained access control for users and hosts to achieve greater storage service levelsSecurely separates data and eliminates unauthorized or accidental accessUp to 1024 Domains with individual settings per 3PAR arrayOptionally Priority Optimization allows assigning QoS to Virtual Domains What are the

48、 benefits of Virtual Domains?36Physical StorageConsolidated StorageCentralized AdminFull set-up, provisioning and monitoring End UsersConsumers onlyProvisionedStorageEnd UserSelf-provisioningVirtual DomainsSecure virtual arrayCentralized Storage Administrationwith Traditional StorageSelf-Service Sto

49、rage Administrationwith 3PAR Virtual DomainsConsolidated StorageCentralized AdminVirtual Domain set-up and monitoring only3PARPriority Optimization373PAR Priority OptimizationStorage based QoSAllows customers to ensure quality of service and better utilize storage resources per Virtual Volume and/ o

50、r Virtual Domain by Setting Priority levels (High, Normal, Low) throttle order for object Setting Service LevelsMax limit IOPS or bandwidth per objectMin goal Min floor for IOPS or bandwidthLatency goal Service level target for an objectCan be enabled, disabled or modified in real time from GUI or C

51、LINo host agents are requiredNo physical partitioning of resources within the storage array is requiredSupports multi-tenant environments383PAR Priority Optimization QoSSimple to implement/monitor/troubleshoot.Requires no physical partitioning.Satisfies a wide set of use cases.Works on any tier of s

52、torage and connectivity protocol. Allows individual Latency, Bandwidth and IOPS settings.Works the same on all 3PAR platforms, from the 7200 to the 20850 No Host Agents required.39Priority Optimization Data Sheet/v2/GetPDF.aspx%2F4AA4-6961ENW.pdfPriority Optimization White Paper /h20195/v2/GetDocume

53、nt.aspx?docname=4AA4-7604ENWStorage Quality of Service (QoS)Guarantee performance with volume-level granularity40Limit bandwidth and/or IOPS on a volume, volume set or Virtual Domain to prevent noisy neighborsAvailable for IOPS, bandwidth and latency, set goals for the system to achieve where possib

54、leFor the ultimate simplicity, set up to three levels of priority: high, medium and lowHighMediumLowImplement limitsSet goalsPrioritize3PAR StoreServ sized for 100000 IOPS Virtual Domain 1 Priority = Highmax IOPS 80000Latency Goal = 8msVirtual Domain 2 Priority = Normalmax IOPS 40000No Latency goal3

55、PAR Leadership Priority Optimization part of Data Optimization Suite Protect your mission critical applications Assure tenant/application level QoS by assigning priorities, targets and caps to VV sets and/or Virtual DomainsMax limit IOPS or bandwidth per objectMin goal Min floor for IOPS or bandwidt

56、hLatency goal Service level target for an objectPriority level throttle order for object 41Read the Evaluator Test Validation: /V2/GetDocument.aspx?docname=4AA5-6504ENW&cc=us&lc=en Tenant 2 ESX ClusterTenant 1 Oracle DB VVset 1 Values inherited from DomainOracleVVset 2 Min 8000 IOPSMax 20000 IOPSPri

57、ority = NormalVM 1VM 2VVset 3Min IOPS 6000, Max IOPS 15000Priority = NormalVM3VVset 4Min 5000 IOPS Max 18000 IOPS Priority = Low3PAR Priority OptimizationBasic max IOPS cappingExampleApplication red is critical and is being under-servicedApplication blue is well behavedApplication green is taking mo

58、re resources than it should and starving red42Enabling QoSBy throttling the VV-set of application green to 4000 IOPS resources for application red are freed-up IOPSTime200060004000100008000264108IO Latency (ms)TimeQoS enabled3PAR Priority OptimizationDynamic capping based on busyness levels and late

59、ncies43IOPSGold (High)Silver (Medium)Bronze (Low)Max 8kMax 6kMax 10kMin goal 5kMin goal 4kMin goal 3k0 25% 50% 75% 100% 10% SLA Bronze setqos -pri low -io 1000-5000 Will be delayed first, not go beyond 5000 IOPS and not delay/reject below 1000 IOPSSLA Silversetqos -pri normal -io 2000-7000 Will be d

60、elayed second, not go beyond 7000 IOPS and not delay/reject below 2000 IOPSSLA Goldsetqos -pri high -io 10000-10000 -lt 10Allows up to 10,000 IOPS and will try to meet the 10ms service times by throttling Bronze and SilverBusy LevelGo back to SW and FeaturesGENERALDashboardActivityScheduleSettingsBL