HP Microsoft Windows Server 2003 for Itanium-based Systems Alternative fault-tolerance methods

Page 65

Item

RAID 0

RAID 1+0

RAID 5

RAID 6 (ADG)

 

 

 

 

 

Tolerates simultaneous failure of

No

Only if no two

No

Yes

more than one physical drive

 

failed drives are

 

 

 

 

in the same

 

 

 

 

mirrored pair

 

 

Read performance

High

High

High

High

 

 

 

 

 

Write performance

High

Medium

Low

Low

 

 

 

 

 

Relative cost

Low

High

Medium

Medium

 

 

 

 

 

*Values for the fraction of drive space usable are calculated with these assumptions: (1) all physical drives in the array have the same capacity; (2) online spares are not used; (3) no more than 14 physical drives are used per array for RAID 5; and (4) no more than 56 drives are used with RAID 6 (ADG).

Selecting a RAID method

NOTE: Not all controllers support RAID 6 (ADG).

Most important criterion

Also important

Suggested RAID level

 

 

 

 

Fault tolerance

Cost effectiveness

RAID 6

(ADG)

 

I/O performance

RAID 1+0

 

 

 

 

Cost effectiveness

Fault tolerance

RAID 6

(ADG)

 

I/O performance

RAID 5

(RAID 0 if fault tolerance is not required)

 

 

 

 

I/O performance

Cost effectiveness

RAID 5

(RAID 0 if fault tolerance is not required)

 

Fault tolerance

RAID 1+0

 

 

 

 

Alternative fault-tolerance methods

Your operating system may also support software-based RAID or controller duplexing.

Software-based RAID resembles hardware-based RAID, except that the operating system works with logical drives as if they were physical drives. To protect against data loss caused by physical drive failure, each logical drive must be in a different array from the others.

Controller duplexing uses two identical controllers with independent, identical sets of drives containing identical data. In the unlikely event of a controller failure, the remaining controller and drives will service all requests.

Neither of these alternative fault-tolerance methods supports online spares or automatic data recovery, nor do they support auto-reliability monitoring or interim data recovery.

If you decide to use one of these alternative methods, configure your arrays with RAID 0 for maximum storage capacity and refer to your operating system documentation for further implementation details.

Drive arrays and fault-tolerance methods 65

Image 65
Contents HP Array Configuration Utility User Guide Audience assumptions Contents Diagnosing array problems Installing ACU Features and system requirementsGetting started This sectionComparison of ACU execution modes Overview for using ACUSetting the execution mode for Microsoft Windows Choosing an operating mode Opening ACU in Local Application modeHttp//servername2301 Opening ACU in Browser modeSelect Device Queries Opening ACU through Systems Insight ManagerGUI operating modes Typical Standard mode screenTypical Configuration Wizards mode screen Typical Express Configuration mode screen Completing the configuration processPossible tasks in Standard Configuration mode Configuring a new controllerUsing Standard Configuration mode Using Express Configuration mode Creating an array Using the configuration wizardsClick Finish Click Create an array, and then click BeginConfiguring a new controller Click Create a logical drive, and then click Begin Creating a logical driveConfiguring a new controller Choices available after opening ACU Modifying an existing configurationModifying a configuration using Standard Configuration mode Modifying a configuration using the Configuration wizards Modifying a configuration using Express modeClear Configuration Click Express Configuration, and then click BeginCreate an array Controller settingsClick Controller Settings, and then click Begin Create a logical drive Delete arrays Click Delete logical drives, and then click Begin Delete logical drivesExpand Array Migrate a logical drive Extend logical driveClick Extend logical drive, and then click Begin Click Migrate a logical drive, and then click BeginSelective Storage Presentation Spare managementRA4x00 controllers Click Spare Management, and then click BeginMSA and Smart Array Cluster storage controllers Click ACU Switch Configuration Configuring switchesRecombining a split, mirrored array Splitting a mirrored arrayModifying an existing configuration Scripting in ACU Command line syntaxIntroduction to scripting in ACU Operating modesCpqacuxe -i Filename Sample custom input scriptDescription of option categories in ACU scripting Script file optionsController category Control categoryAction MethodController ClearConfigurationWithDataLossLicenseKey, DeleteLicenseKey RAIDArrayIDArray category SurfaceScanDelayRebuildPriority, ExpandPriority SSPStateOnlineSpare Logical Drive categoryDrive LogicalDriveSectors SizeStripeSize ArrayAcceleratorSSPAdaptersWithAccess Error reportingConnectionName HostModeACU scripting error messages 0-9, !, @, #, *, , , ,, -, , +, , ., /, and space. The ID Current configuration Running the CLI Using the Command Line InterfaceOverview of the ACU CLI CLI syntax Command variableKeyword abbreviations Target variableExample command Typical proceduresHiding warning prompts Querying a deviceSyntax Sample scenarioExample Example commandsModifying the controller chassis name Sample scenario Using Selective Storage PresentationManaging host modes connection profiles Modifying connection namesIdentifying devices Deleting target devicesExpanding an array Extending a logical driveManaging spare drives Target modify raid=01+0156adg? ss=8163264128256default?Migrating a logical drive Target modify size=#max? forcedChanging the controller cache ratio Changing the Rebuild Priority settingChanging the Expand Priority setting Enabling or disabling the drive cache Changing the surface scan delay timeRe-enabling a failed logical drive Enabling or disabling the array accelerator Setting the targetDisabling a redundant controller Typical scenario= rescan Rescanning the systemFactors involved in logical drive failure Probability of logical drive failureProbability of logical drive failure Drive arrays Drive arrays and fault-tolerance methodsDrive arrays and fault-tolerance methods Hardware-based fault-tolerance methods Fault-tolerance methodsDisadvantages RAID 0-No fault toleranceAdvantages RAID 1+0-drive mirroring RAID 6 ADG-Advanced Data Guarding RAID 5-distributed data guardingRAID RAID 6 ADG Comparing the hardware-based RAID methodsSelecting a RAID method Alternative fault-tolerance methodsDiagnostic tools Diagnosing array problemsServer Diagnostics Post messagesAcronyms and abbreviations SSP Index Index