International Business Machines Corporation

James Christopher Wyllie

Frank Bernhard Schmuck

Roger Lee Haskin

Robert J Curran

Wayne A Sawdon

Zvi Yosef Yehudai

Veera W Deenadhayalan

John Michael Marberg

Irit Loy

Marc M Eshel

Boaz Shmueli

Ralph Attila Becker-Szendy

Daniel Lloyd McNabb

Boaz Shumeli

Dean Hildebrand

David Jones Craft

D Scott Guthridge

James Lee Hafner

Wendy Ann Belluomini

Yuri L Volobuev

Benjamin Mandler

Brian P Dixon

Sibylle Schaller

Jaishankar Moothedath Menon

William A Kish

Ashok Raj Saxena

Kuei-Yu Wang-Knop

Anthony J Zlotek

Radha R Kandadai

Thomas Eugene Engelsiepen

Martha R Henley

Geoffrey Albert Riegel

Scott Lyon Porter

Scott D Guthridge

Demetrios K Michalaros

Carol D Hartman

Daniel L Mcnabb

Scott L Porter

Veera c/o Ibm Uk Limited Intellectual Property Law Deenadhayalan

Embodiments of the invention relate to supporting coordinated access to a clustered file system's shared disk storage subsystem by using dynamic creation of file access layout for different workloads and access patterns. Active file servers in the cluster have equivalent access to shared data of said file system. In one embodiment, a particular file server in the cluster is configured to coordinate the file system's layout. The particular file server, in response to a particular client's request for a layout of a file stored by said file system in said subsystem determines a file access layout for a parallel file access protocol to use when servicing an I/O request by said client, marks said layout as active for the client, and sends the layout to the client. The file access layout is determined based on one or more of current file system resources, a specific workload type and a specific byte-range of the file included in the client request, and properties of the file. The layout includes a set of file servers in said cluster and byte ranges of the file, and the byte-range being associated with a particular server among the set.

Coordinated access to a file system's shared storage using dynamic creation of file access layout

A method for performing stripe placement within a storage system is disclosed. After a set of failure domains within a storage system has been identified, the failure domains are then organized to form a hierarchy of failure domains. A failure domain is defined as a group of one or more disks that are more likely to fail together because a common component is shared by that group of disks. Stripe placement is performed across all active failure domains within the storage system using a greedy algorithm.

Failure domain based storage system data stripe layout

Embodiments of the invention relate to coordinated access to a clustered file system's shared storage subsystem using a shared-lock architecture. A particular file server is configured to coordinate the file system's file access layout particular server, in response to a client's file access layout request for byte-ranges of a file stored in the storage subsystem being received by a file server in cluster and sent to said particular server, generates a file access layout of a set of file servers for a parallel file access protocol to use and byte ranges of the file that particular individual file servers to use, to service an I/O request for the file. The file server (that received the request), sends the generated file access layout to said client, and requests coordinated access to said file from the particular server in response to receiving an I/O request to byte-ranges of said file.

Coordinated access to a clustered file system's shared storage using shared-lock architecture

Embodiments of the invention relate to accessing data that spans multiple storage devices in a storage system. A first index records identifiers of data blocks and their storage locations, and a second index records storage locations. An update request is processed by writing multiple copies of a received data block identifier and the updated data block to storage locations recorded in the second index. A read request is processed by performing a quorum read of the storage locations recorded in the second index to retrieve the updated data blocks. The processing of the read request also includes performing a quorum read of the first index to determine storage locations that contain data blocks to be read as part of processing the read request. The data blocks to be read exclude the updated data blocks.

Accessing data in a storage system

Embodiments of the invention relate to supporting coordinated access to a file system's shared disk storage subsystem using an automatic alignment of a parallel file access protocol and metadata management on a per object basis. A file server (among servers in the cluster) receives a client's request to access a file stored by the file system in the storage subsystem using a parallel file access protocol. If the file system has selected a file server to serve as a metadata manager of said file, the file server uses the selected file server to coordinate access to the file's metadata. If the file system has not selected a file server to serve as the metadata manager of the file, the file system selects the file server that received said request as the metadata manager for said file until a file close request operation is received.

Supporting coordinated access to a file system's shared storage using automatic alignment of a parallel file access protocol and metadata management

A data validation method comprises associating a validation chunk with one or more data chunks stored on one or more storage mediums. The validation chunk comprises validation information for verifying whether data recorded in a subdivision of the one or more data chunks is lost. A request is received for an operation to be performed on first data stored in a first subdivision of a first data chunk from among said one or more data chunks. One or more portions of the validation information are examined to determine whether the data recorded in said first subdivision is deemed to be lost.

Monitoring lost data in a storage system

This disclosure relates to parity declustered storage device arrays having partition groups. In an exemplary embodiment, the storage system includes a storage device array, such as disk array. Each storage device is divided into partitions. Each partition includes stripe units, such as hundreds or thousands of stripe units in exemplary embodiments. The storage system also includes a physical array controller coupled to the storage device array. In an exemplary embodiment, the array controller includes a partition group lookup table and stores and retrieves data and parity in the storage devices based on the partition group lookup table. In this exemplary embodiment, the array controller also includes a stripe lookup table and/or a log. In an exemplary embodiment, the partition group lookup table and the stripe lookup table take up less memory (e.g., by an order of magnitude) than a single-level stripe map conveying the same information.

Parity declustered storage device array with partition groups

Data validation systems and methods are provided. Data is recorded in N data chunks on one or more storage mediums. A first validation chunk independently associated with said N data chunks comprises first validation information for verifying accuracy of data recorded in said N data chunks. The first validation chunk is associated with a first validation appendix comprising second validation information, wherein the first validation appendix is stored on a first storage medium independent of said one or more storage mediums.

Data integrity validation in storage systems

Space is allocated on data storage devices in proportion to weights associated with the storage devices. The weights can be dynamically adjusted at any time in order to accommodate changes in the system and to better utilize the storage devices. The technique used to perform the allocating is independent of the weights used by the allocating. Further, the allocation technique can accommodate general purpose data streams having varying lengths and/or varying access patterns, as well as special purpose data streams, such as video streams.

Allocating space on data storage devices in proportion to weights associated with the devices

The decision to prefetch inodes is based upon the detecting of access patterns that would benefit from such a prefetch. Once the decision to prefetch is made, a plurality of inodes are prefetched in parallel. Further, the prefetching of inodes is paced, such that the prefetching substantially matches the speed at which an application requests inodes.

Detecting when to prefetch data and then prefetching data in parallel

A system for balancing multiple memory buffer sizes includes a memory partitioned into a plurality of subgroups, each subgroup containing similarly sized buffers. The system further includes a memory manager which manages requests for buffer space by storing and accessing buffer steal history records in a buffer steal history queue.

System for balancing multiple memory buffer sizes and method therefor

Detecting when to prefetch inodes and then prefetching inodes in parallel

Growing community of inventors

Monte Sereno, CA, United States of America

James Christopher Wyllie