International Business Machines Corporation

Globalfoundries Inc.

Alastair Cooper

Gordon Douglas Hutchison

John Paul Agombar

Ian Boden

Robert Bruce Nicholson

Nicholas Michael O'Rourke

Dominic Tomkins

Gemma Izen

Kurt Allen Lovrien

David Ray Kahler

Michael R Groseclose, Jr

Mark Keith Elliott

Daniel Paul Dent

Method and system are provided for dynamic rebuild capability in redundant array of independent disks (RAID) arrays using compressing drives. The method includes providing an array including a physical rebuild area for the multiple drives of the array and dynamically adjusting a number of allocated rebuild zones available within the rebuild area, wherein each allocated rebuild zone has capacity to store a drive rebuild based on a current physical usage of the multiple drives of the array.

Dynamic rebuild capability in redundant array of independent disks (RAID) arrays using compressing drives

Operating a data storage system comprising a plurality of disk drives and a storage controller connected to the disk drives. A first subset and a second subset of the plurality of disk drives are operated as short stroked disk drives and non-short stroked disk drives, respectively. Priority storage spaces are defined including a high priority storage space, a medium priority storage space, and a low priority storage space. Data is received including associated access rates for each portion of the data. One of the priority storage spaces is identified to store a portion of the data, based on the access rates for each portion of the data. Data accessed most frequently is stored in the high priority storage space, data accessed least frequently is stored in the low priority storage space, and the remaining data is stored in the medium priority storage space.

Controlling operation of a data storage system

A computer-implemented method for performing garbage collection in a data storage system having a RAID configuration according to one embodiment includes determining on which geometry addressable region of the RAID configuration extents of data are stored. The RAID configuration includes an old geometry addressable region and a new geometry addressable region. The method further includes determining an order for performing garbage collection on the extents of data. Extents of data determined to be stored on the old geometry addressable region are assigned a higher priority within the order than extents of data determined to be stored on the new geometry addressable region. Moreover, the method includes performing garbage collection on the data storage system according to the order for performing garbage collection.

Selective use of garbage collection during expansion of geometry addressable regions of a redundant array of independent drive (RAID) configuration

Method and system are provided for handling an underperforming drive in a redundant array of independent disks (RAID) operating a distributed sparing mechanism. The method includes marking an underperforming drive in the array as resting and allocating distributed spare space of the array for data of operations to the resting drive. A data structure is designated to track operations diverted to the distributed spare space from the resting drive. If the resting drive is recovered, the method rebuilds data to the recovered drive from the distributed spare space by reference to the data structure.

Underperforming drive handling in redundant arrays

A RAID controller may update a RAID array by receiving updated data for a first data strip in a set of data strips in the RAID array. The RAID controller may then determine that the first data strip is stored on a device that is experiencing a slow condition. The RAID controller may then force, based on the determining, a promoted stripe write.

Update of RAID array parity

Operating a data storage system comprising a plurality of disk drives and a storage controller connected to the disk drives. A first subset and a second subset of the plurality of disk drives are operated as short stroked disk drives and non-short stroked disk drives, respectively. Priority storage spaces are defined including a high priority storage space, a moderate priority storage space, and a low priority storage space. Data is received including associated access rates for each portion of the data. One of the priority storage spaces is identified to store a portion of the data, based on the access rates for each portion of the data. Data accessed most frequently is stored in the high priority storage space, data accessed least frequently is stored in the low priority storage space, and the remaining data is stored in the moderate priority storage space.

A method for minimizing head seek movement and improving I/O performance of a hard disk drive is disclosed. In one embodiment, such a method includes logically dividing storage space of a hard disk drive into storage areas of substantially equal size. The method monitors a temperature of each of the storage areas. The temperature indicates how frequently data in a corresponding storage area is accessed. The method swaps data in storage areas of the hard disk drive based on temperature. These swaps involve moving hotter data toward outer tracks of the disk drive and colder data toward inner tracks of the disk drive. A corresponding system and computer program product are also disclosed.

Temperature-based disk defragmentation

A method for balancing performance in a redundant array of independent disks (RAID) made up of heterogeneous storage drives is disclosed. In one embodiment, such a method includes implementing a RAID on a set of storage drives. The set of storage drives includes a lower-performing storage drive that underperforms other storage drives in the set. In order to balance performance in the RAID, the method stripes data across the set of storage drives such that the lower-performing storage drive participates in only a selected percentage of strides. In certain embodiments, this percentage is selected based on an amount the lower-performing storage drive underperforms other storage drives in the set. A corresponding system and computer program product are also disclosed.

Performance-balanced heterogeneous raid

A method for recovering a failed storage drive in a redundant array of independent disks (RAID) includes storing, on a first distributed spare of the RAID, a first parity subset for a first set of drives in the RAID. The method further stores, on a second distributed spare of the RAID, a second parity subset for a second set of drives in the RAID. In the event a storage drive in the RAID fails, the method determines whether the storage drive belongs to the first set or second set. In the event the failed storage drive belongs to the first set, the method uses the first parity subset on the first set of drives in the RAID to recover the failed storage drive. In the event the failed storage drive belongs to the second set, the method uses the second parity subset on the second set of drives in the RAID to recover the failed storage drive. A corresponding system and computer program product are also disclosed.

Disposable subset parities for use in a distributed RAID

Methods, apparatus and computer programs have been provided for mitigating a problem of non-optimal recovery from storage device failures. A method involves determining a required write performance for rebuilding data of a failed device, based at least partly on the potential read performance of storage devices in a data rebuild; and allocating a virtual storage area within available storage, which allocation of virtual storage is based at least partly on the required write performance. Data is rebuilt by writing to the allocated virtual storage area, and data recovery is completed by migrating this rebuilt data to at least one data storage device such as a spare physical storage device within the array.

Controlling data storage in an array of storage devices

Exemplary embodiments of the present invention disclose a method and system for reducing a probability of generating an unrecoverable error on a disk array during a disk rebuild. In a step, an exemplary embodiment identifies a disk to be replaced in the disk array, the disk array including a spare disk. In another step, an exemplary embodiment locates a region in the disk array that incurs a high number of reads and writes during a period prior to replacing the disk in the disk array. In another step, an exemplary embodiment scrubs data in a region in the disk array that has incurred a high number of accesses. In another step, an exemplary embodiment replaces the disk identified to be replaced with the spare disk in the disk array. In another step, an exemplary embodiment rebuilds data on the replaced disk on the spare disk in the disk array.

Preventing unrecoverable errors during a disk regeneration in a disk array

Managing power-consumption, for use in a storage system comprising first data stored on one or more storage devices is provided. A receiver receives a policy comprising a power-management parameter, a first parameter and a rule associated with the policy. An analyser, responsive to receipt of a policy, analyzes second data associated with the rule. A determiner, responsive to the analysis, determines third data in accordance with the second data. The third data is associated with migration of the first data.

Growing community of inventors

Winchester, United Kingdom

Alastair Cooper