Broadcom Corporation

Adaptec, Inc.

Digital Equipment Corporation

Raidcore, Inc.

Jeffrey T Wong

Chris Robert Franklin

Randy M Arnott

Eric S Noya

Stanley Luke

Timothy Lee Hoskins

Bryan K Panner

Richard Napolitano

Paul Richard Shaughnessy

Chester Juzsczak

Alexander C McLeod

Herbert W Silverman

Mitchell N Rosich

John F Considine

Angela E Overman

A method, system and computer program product for a generic data storage interface for local and remote networked storage is provided. It comprises providing a data storage interface accessible by an operating system, transmitting data from an application running on the operating system to the data storage interface and selectively storing the data using the same data storage interface in either a local or a remote networked storage location. Firmware for the common data storage interface assigns a unique identifier that classifies an I/O request as a local or remote networked I/O request and adds the I/O request to a common stack for local and remote networked data storage elements. When an I/O request is removed from the stack, the unique identifier is used to identify the I/O request as a local or remote networked I/O request. Local I/O requests are executed via a local bus such as PCI, or SCSI. Remote networked I/O requests require a connection to one or more remote networked storage elements via a network. Memory space in a cache may be allocated for both local or remote networked I/O requests. DMA transfer may be employed for both local and remote networked I/O requests. Both local and remote networked I/O requests are executed using the same firmware and the same data interface.

Combined local and network storage interface

Self-generated automated tests can use a pseudo-random number generator to select one or more arguments that are passed to programs and scripts. The random arguments are driven by a configuration file where the limits for the parameters are defined. Multidimensional functions with multidimensional parameters can be tested. Test duration can be limited by time, number of iterations, or by any of the multidimensional functions or parameters. A pseudo-random seed for each test is recorded so that a test case can be reproduced if a failure is detected or otherwise.

Self-generated test automation

A method and related computer program product of preventing write corruption in a redundant array in a computer system, comprising detecting a write failure from a calling application to at least one disk of the redundant array, writing failure information to non-volatile storage; returning an I/O error to the calling application; reading the failure information from the non-volatile storage during the next system reboot; and reconfiguring the array to eliminate the failed disk.

Method and computer program product for marking errors in BIOS on a RAID controller

A method and related computer program product for migrating legacy data to a RAID array while contemporaneously providing user access to legacy data, comprising connecting the legacy drive to the RAID controller, converting the legacy drive into a legacy array, selecting the legacy array and selecting the RAID array to migrate the legacy data to.

Method and computer program product to migrate legacy data to a RAID array while contemporaneously providing user access to legacy data

A method and related computer program product for storing first configuration information for a plurality of logical devices coupled to a RAID controller. Subsequent configuration information is stored for the plurality of logical devices coupled to the RAID controller while retaining previously written configuration information. Finally, in the event of a conflict in configuration information, the first and subsequent configuration information is compared to determine the cause of the conflict.

Method and computer program product of keeping configuration data history using duplicated ring buffers

A finite state machine (FSM) for a redundant array of independent disk includes a single process context that maintains an entire finite state required for input/output operations performed in a RAID system. The finite state is only updated in response to calls and call-backs. The call-backs can include procedure returns and interrupt signals. The call is received directly from an application program, and the call-backs are received from a driver and passed back directly to the application software by the finite state machine.

Finite state machine with a single process context for a RAID system

A fault tolerant method transforms physically contiguous data in-place on a disk by partitioning the physically contiguous data into an empty region physically adjacent to data regions including a first data region and a last data region, the first and last data regions at opposing ends of the physically contiguous data regions. The physically contiguous data are transformed in an order beginning with the first data region and ending with the last data region. The transforming step perform first locking and reading the first data region, second, transforming the first data region, third, writing and unlocking the transformed first data region to the empty region, and fourth, declaring the first data region as the empty region while declaring the empty region as the first region. The first through fourth steps are repeated for each data region, until completion, to transform the physically contiguous data in-place on the disk.

In-place data transformation for fault-tolerant disk storage systems

A method to expand a RAID subsystem from a first array of disk drives to a second array of disk drives. The first array includes a set of data disk drives storing old data and spare space, and the second array includes the first array and at least one new disk drive. First, the old data are distributed among the set of data disk drives and at least one new disk drive while, at the same time, new data are mapped to the spare space. Upon completion of the distribution, the new data are copied from the spare space to the set of data disk drives and at least one new disk drive to enable concurrent expansion of the first array while accessing the old and the new data.

Expansion of RAID subsystems using spare space with immediate access to new space

A method and system that allows the distribution of hot spare space across multiple disk drives that also store the data and redundant data in a fully active array of redundant independent disks, so that an automatic rebuilding of the array to an array of the identical level of redundancy can be achieved with fewer disk drives. The method configures the array with D disk drives of B physical blocks each. N user data and redundant data blocks are allocated to each disk drive, and F free blocks are allocated as hot spare space to each disk drive, where N+F&#x3c;=B, and ((D&#x2212;M)&#xd7;F)&#x3e;=N. Thus, rebuilding of data and redundant blocks of a failed disk drive in the free blocks of the remaining disk drives is enabled after M disk drive failures.

Rebuilding redundant disk arrays using distributed hot spare space

A configurable RAID subsystem includes a user data array connected to a user application via a block I/O path, and a configuration array connected to a configuration application via the same block I/O path. The user data array processes user data access commands executed by the user application; and the configuration application processes configuration commands, the user data access commands and the configuration commands communicated to the user data array and the configuration array respectively, via the block I/O path. A dynamic identification is assigned to the user data array by the configuration array, and a static identification is assigned to the configuration array.

Method and system for configuring RAID subsystems with block I/O commands and block I/O path

A system and method for updating parity based upon locking and unlocking of a storage stripe in a redundant array of independent disk (RAID) implementation is provided. The stripe includes a parity block. The stripe is locked to prevent colliding I/O operations from being performed thereto while a current I/O operation is underway with respect to the stripe. A parity buffer is maintained that is updated to include the current parity information for the stripe. The buffer is &ldquo;swapped&rdquo; with the parity buffer associated with a next waiting I/O operation request before the stripe is unlocked. The buffer continues to be swapped with further requests so long as another I/O operation request waits on the lock. When no further I/O operation request for the given stripe is detected, then the current parity buffer is written into the stripe parity block. The intervening swaps reduce the number of parity cache reads and writes, increasing efficiency.

System and method for parity caching based on stripe locking in raid data storage

A system and method for changing between various levels and configurations of redundant array of independent disk (RAID) implementations is provided. A specialized driver establishes a specialized &ldquo;morph&rdquo; container in a top level above primary and secondary level containers. The morph container communicates with the host computer I/O and arranges mapping of data between an original source container configuration and a new destination configuration. A morph container mapping structure is implemented to accomplish the mapping therebetween. Where data is migrated from an original container space back into an original space, a temporary container can be established to facilitate the transfer between the space.

System and method for accomplishing data storage migration between raid levels

A system and method for rapidly zeroing/clearing a container in a redundant array of independent disks (RAID) provides the writing of a series of logical zeros to each disk in the container using an internal bus driver-level command. The command causes a small data block of all zeroes written to the disk to be duplicated so that the entire container space in each disk is effectively written-to. The bus driver is preferably a small computer system interface (SCSI) architecture that supports a WRITE SAME command. Where the disk device fails to support the WRITE SAME command then an alternative process is employed, in which a single large sized memory block is created and initial sized with all zeroes. A predetermined number of virtual scatter gather elements are created, each pointing to the memory block. The data in these scatter gather elements is written to the unsupported device so as to clear the entire container space of that device using a smaller block than the overall space requires for complete clearance.

System and method for zeroing data storage blocks in a raid storage implementation

A file array storage architecture comprises a file system that is distributed across a data processing platform consisting of a host computer coupled to an adapter over a high-speed, low latency interface and an input/output subsystem that is located entirely on the adapter. The distributed nature of the architecture enables implementation of the file system in accordance with a modified client-server computing model; that is, the architecture includes a client file system executing on the host computer and a server file system executing on the adapter.

Growing community of inventors

Newton, MA, United States of America

Jeffrey T Wong