Dot Hill Systems 200 manual Features, Controller Optimization, Controller Redundancy

Page 14

Features

Controller Optimization

SANnet II 200 SCSI array controller operations are highly automated to minimize management workloads and are self-tuning to optimize performance. Powerful advanced features are supported, such as multi-threading, segmentation, read-ahead, write-through, write-back, queuing and sorting. Array users can adjust controllers to more closely match their particular environment by selecting between optimization for random or sequential access patterns.

Array controllers segment their cache into a set of relatively small cache allocation units. The unit capacity depends on the user-selectable cache optimization setting. Whenever the random option is selected, the cache is divided into units of 4KB each. Units of 16KB are created when the sequential option is selected. A cache unit is fully independent of the others and can serve any one logical drive at a given moment.

Each unit is automatically allocated by the controller’s caching algorithms to provide read caching or write caching and dynamically reassigned as needed. For example, when the number of vacant cache units becomes low, write-back units might be flushed to disks or read-ahead units may be unallocated to make them available for other uses. Empty cache units can be allocated for any necessary purpose, regardless of the assignment they most recently held.

A controller’s optimization setting also affects how logical drives are organized and initialized. When the random option is selected, each disk used in a logical drive will be segmented into 32KB per segment. When the sequential option is selected, the segment size is 128KB. The sequential optimization option must be used when creating logical drives with total storage capacities of 512GB or more.

The controller optimization setting has a profound effect on the logical organization and management of cache and disks. An option must be selected before logical drives can be created, but the optimization setting cannot be changed dynamically. For maximum flexibility, choose the sequential optimization option to allow the creation of logical drives of any size.

Controller Redundancy

SANnet II 200 SCSI arrays are available with single or redundant controllers. When two controllers are present, they always operate as a redundant pair. They cannot function as two individual, i.e., single controllers. Redundant controllers can be set to function in active/active (recommended) or active/standby pairs.

Unlike other storage systems, redundant controllers behave as one virtual system consisting of primary and secondary controllers. Configuration, management and monitoring are performed using the primary controller only. The secondary controller automatically synchronizes its configuration to the primary controller to ensure their configuration remains identical.

If an active controller detects the other controller of a redundant pair is malfunctioning or has failed, it immediately disables the faulty controller and transparently resumes its workload. This temporary process, called failover, occurs almost immediately and without interuption to server or array operations.

14

Image 14
Contents SANnet II 200 Scsi Array Technical Product Description Guide Copyright Contents Key Features IntroductionArchitecture Fundamentals Management InterfacesReliability, Availability and Serviceability Typical Sequence of Installation StepsPerformance ScalabilityController Modules ArchitectureEnclosure Modules SANnet II 200 Scsi Array RAID Controller Diagram Disk ModulesPower and Cooling Modules Disk Drives front view with bezel removedEnclosure Management Unit Modules Enclosure Management UnitScsi Module Scsi Module i.e., I/O ModuleTermination Module Termination ModuleController Redundancy FeaturesController Optimization Cache Protection Read CachingDisk Redundancy Write CachingLogical Drives Spare Disks Logical Drive RebuildingConfiguration ManagementMonitoring Event LogsEvent Log Window In-Band & Out-of-Band Management In-Band ManagementTypical Diagnostic Service and SANscape Setup Event ReportingFirmware Updates Detailed Specifications CachingRAID3 Advanced FeaturesRAID5 RAID Expansion Redundant ControllersData Safety User Interface JBOD, N-RAID, RAID 0, RAID 1, RAID 1+0, RAID 3, RAID Performance ParametersParameter Range Cispr GlossaryAnsi CLIGbic FC-ALFRU HBAJbod IECIops LANRAID NvramPID RASSES SCAScsi SIDWWN XOR