International Business Machines Corporation

John Earle Lindley

William James Scales

Carlos Francisco Fuente

Kenneth Wayne Boyd

John Paul Agombar

Lee Jason Sanders

Matthew John Fairhurst

Gregory E McBride

Kenneth Fairclough Day, Iii

Eric John Bartlett

Kiron Haltore

Colin R Jewell

Tu-An Cheng

Yasumitsu Mizoguchi

Kenta Ninose

Akinobu Shimada

Tsutomu Sukigara

Jialin Ju

Gerardo De Jesus Frias Garcia

Ryusuke Ito

Aviad Zlotnick

Michael Edward Factor

Robert Bruce Nicholson

Jaishankar Moothedath Menon

Shachar Fienblit

Orit Nissan-Messing

Thomas Charles Jarvis

Olympia Gluck

Kenneth F Day, Iii

Robert Wesley Shomler

Douglas Paul Dunham

Clement Lambert Dickey

Nancy Lott Hobbs

Tina Lynn Dunton

Barbara J Camacho

Roman David Druker

An apparatus, method and computer program in a distributed cluster storage network comprises storage control nodes to write data to storage on request from a host; a forwarding layer at a first node to forward data to a second node; a buffer controller at each node to allocate buffers for data to be written; and a communication link between the buffer controller and the forwarding layer at each node to communicate a constrained or unconstrained status indicator of the buffer resource to the forwarding layer. A mode selector selects a constrained mode of operation requiring allocation of buffer resource at the second node and communication of the allocation before the first node can allocate buffers and forward data, or an unconstrained mode of operation granting use of a predetermined resource credit provided by the second to the first node and permitting forwarding of a write request with data.

Dynamic resource allocation for distributed cluster-storage network

A method for operating a mirrored disk storage system, comprises a determiner component for determining a repair characteristic of a pair of mirrored disks, a comparator for comparing said repair characteristic of a first of said pair with said repair characteristic of a second of said pair, and a cancelling component for cancelling a repair action for said one of said first or said second of said pair that is not selected by said selector.

Recovery control in mirrored disks

A method according to one embodiment includes receiving an I/O request for data, and determining whether the I/O request refers to migrated or unmigrated data. Unmigrated data is data residing on a first portion of a drive array, the first portion being arranged in according to a first array configuration having at least a first parameter. Migrated data is data that has been migrated from the first portion to a second portion of the drive array, the second portion having a second array configuration having at least a second parameter. The method also includes using parameters corresponding to the first array configuration in response to determining that the requested data is unmigrated data. However, in response to determining that the requested data is migrated data, parameters corresponding to the second array configuration are used.

Raid array transformation in a pooled storage system

An apparatus and computer program in a distributed cluster storage network comprises storage control nodes to write data to storage on request from a host; a forwarding layer at a first node to forward data to a second node; a buffer controller at each node to allocate buffers for data to be written; and a communication link between the buffer controller and the forwarding layer at each node to communicate a constrained or unconstrained status indicator of the buffer resource to the forwarding layer. A mode selector selects a constrained mode of operation requiring allocation of buffer resource at the second node and communication of the allocation before the first node can allocate buffers and forward data, or an unconstrained mode of operation granting use of a predetermined resource credit provided by the second to the first node and permitting forwarding of a write request with data.

Aspects of the present invention relate to an apparatus and corresponding method that includes reading data from a first portion of an in-use redundant array of independent disks (RAID), wherein the first portion of the in-use RAID is arranged according to a first array configuration having at least a first parameter. One or more reserved extents in a storage pool is used to store at least a first portion of the data for writing to the in-use RAID. One or more free extents in the storage pool and/or in the in-use RAID is used to store at least a second portion of the data for writing to the in-use RAID. The data is written to a second portion of the in-use RAID. The second portion of the in-use RAID is arranged according to a second array configuration having at least a second parameter.

RAID array transformation in a pooled storage system

In one embodiment, a method includes reserving extents in a storage pool, reading data from a first portion of an in-use RAID arranged according to a first array configuration, using the reserved extents in the storage pool to store a first portion of the data for writing to the in-use RAID, using one or more free extents in the storage pool or in the in-use RAID to store a second portion of the data for writing to the in-use RAID, writing the data to a second portion of the in-use RAID arranged according to a second array configuration, performing one or more first I/O operations according to the first array configuration, wherein the one or more first I/O operations are performed on the data before the data is read by the data migrating component, and performing one or more second I/O operations according to the second array configuration.

A method for operating a mirrored disk storage system, comprises a detector component operable to detect that a failure has left read stability in doubt, a determiner component for determining a repair characteristic of a pair of mirrored disks, a comparator for comparing the repair characteristic of a first of the pair with the repair characteristic of a second of the pair, and a selector for selecting one of the first or the second of the pair having a preferred repair characteristic to be used as a synchronization source disk. The preferred repair characteristic comprises the characteristic that no repair is required for the one of the first or the second of the pair, or that a least repair time is required for the one of the first or the second of the pair.

An apparatus and system are provided for isolating a storage application from a network interface driver. The apparatus includes a definition module, an application interface, a translation module, and a data buffer. The definition module defines a set of storage application functions. The translation module translates each storage application function to one or more network interface driver functions. The data buffer transfers data blocks between the storage application and the network interface driver.

Apparatus, system, and method for isolating a storage application from a network interface driver

A mechanism is provided for managing operations in a data storage system. A first storage controller monitors operation state information associated with a plurality of operation states stored in a data structure. The first storage controller detects a change in the operation state information. The change in the operation state information is in response to a managing entity updating the operation state information to a first operation state in the plurality of operation states in the data structure. Responsive to detecting the change in the operation state information, the first storage controller identifies a reference to a first operation associated with the first operation state updated by the managing entity. The first operation is one operation in a plurality of operations. The first storage controller performs the first operation associated with the first operation state.

Idempotent storage replication management

Provided are a method, system, and article of manufacture for controlling a first storage system receiving commands from a first and second managers to create a consistency group with a second storage system. Host writes are received at the first storage system, wherein the first storage system includes a first storage system primary site and a first storage system secondary site. The first storage system sends the host writes from the first storage system primary site to the first storage system secondary site. Host write operations are quiesced at the first storage system in response to a first command from a first manager. Host write operations are resumed at the first storage system in response to receiving a second command from the first manager. The first storage system receives a run command with a marker, wherein the marker indicates a cycle number to control the cycles of the first and second storage systems. The first storage system sends the marker from the first storage system primary site to the first storage system secondary site. The first storage system sends the marker to a second manager. The first storage system applies the host writes to the first storage system secondary site. The first storage system sends a first message to the first storage system primary site from the first storage system secondary site after completing the applying of the host writes. The first storage system sends a second message to the first manager indicating whether conformation was received from the first storage system secondary site that the host writes were applied.

Managing consistency groups using heterogeneous replication engines

Provided are a method, system, and article of manufacture for managing consistency groups using heterogeneous replication engines replicating data. A common linkage manager issues a prepare to create consistency group command to a first replication manager and a second replication manager, wherein the first and second replication managers issue create consistency group commands to first and second replication engines, respectively, to form a consistency group of data at a first source storage and a second source storage to copy to a first target storage and second target storage, respectively. The first and second replication managers form a first consistency group and a second consistency group of data from the first and second source storages, respectively, that are consistent as of a point-in-time in response to the create consistency group command; transmitting, by the first and second replication managers, the data in the formed first and second consistency groups to a first journal and second journal, respectively. The first and second replication managers commit the data in the first and second consistency groups in the first and second journals to the first and second target storages, respectively, in response to the first and second replication managers receiving a commit command.

An apparatus, system, and method are provided for isolating a storage application from a network interface driver. The apparatus includes a definition module, an application interface, a translation module, and a data buffer. The definition module defines a set of storage application functions. The translation module translates each storage application function to one or more network interface driver functions. The data buffer transfers data blocks between the storage application and the network interface driver.

A storage area network has one or more digital data processors, e.g., hosts, in communication with one or more storage devices. At least a selected one of the hosts has an operating system in which a storage device must be claimed before the storage device can be accessed by applications programs executing on that host. The improvement is characterized by a selectively actuable filter, e.g., loaded with the selected host operating system, that when downloaded and actuated to a host intervenes to block claiming of one or more selected storage devices.

Storage area network methods and apparatus for dynamically enabled storage device masking

A method for managing cache accessing of CKD formatted records that uses a Predictive Track Table to reduce host delays resulting from cache write misses. Because a significant portion of CKD formatted DASD tracks contain records having no key fields, identical logical and physical cylinder and head (CCHH) fields and similar-sized data fields, a compact description of such records by record count and length data, indexed by track, can be quickly searched to determine the physical track location of a record update that misses the cache. The Predictive Track Table search is much faster than the host wait state imposed by access and search of the DASD to read the missing track into cache. If the updated record that misses cache is found within the set of records in the Predictive Track Table, then the update may be immediately written to cache and to a Non-Volatile Store (NVS) without a DASD read access. This update then may be later destaged asynchronously to the DASD from either the cache or the NVS. Otherwise, if not found in a predictive track, the update record is written directly to the disk and the cache, subject to the LRU/MRU discipline, incurring the normal cache write-miss host wait state.

Growing community of inventors

San Jose, CA, United States of America

John Earle Lindley