International Business Machines Corporation

Ronald E Bretschneider

Beth Ann Peterson

Harry M Yudenfriend

Dale Francis Riedy, Jr

Susan Kay Candelaria

John Richard Paveza

Max D Smith

Gail Andrea Spear

Peter Grimm Sutton

William Stuart Edwards

Lyle LeRoy Merithew, Jr

Marc Kenneth Duquette

Ruth Leslie Ferziger

Peter Verl Gibson

Chung Man Fung

Brian S Lee

Michelle Parks Dais, Legal Representative

Provided are a computer program product, system, and method for managing access requests from a host to tracks in storage. A cursor is set to point to a track in a range of tracks established for sequential accesses. Cache resources are accessed for the cache for tracks in the range of tracks in advance of processing access requests to the range of tracks. Indication is received of a subset of tracks in the range of tracks for subsequent access transactions and a determination is made whether the cursor points to a track in the subset of tracks. The cursor is set to point to a track in the subset of tracks and cache resources are accessed for tracks in the subset of tracks for anticipation of access transactions to tracks in the subset of tracks.

Pre-allocating cache resources for a range of tracks in anticipation of access requests to the range of tracks

Write transfer resource management in a data storage system in accordance with the present description includes overdue write transfer management logic which detects whether or not an established write set has become stale. In one embodiment, a determination is made as a function of whether a write transfer from a host and associated with an established write transfer set is overdue as measured by a time-out period of time. Upon determination that an established write transfer set has become stale, the stale write set is removed and the resources associated with the removed write set are freed for use by other write sets, significantly improving system performance. Other features and aspects may be realized, depending upon the particular application.

Write transfer resource management in a data storage system

Provided are a computer program product, system, and method for processing a write of records to maintain atomicity for writing a defined group of records to multiple tracks. A write is received comprising defined groups of records to write to a plurality of tracks in the primary storage system. Sub-writes are generated to write the records in the defined groups in the write. At least one of the sub-writes includes records to write for at least one of the defined groups that spans multiple tracks in response to determining that at least one of the defined groups includes records that will be written to multiple tracks. The sub-writes are transmitted to the primary storage system to mirror to the secondary storage system to cause the secondary storage system to return complete in response to completing writing to all the tracks written to by the sub-write.

Processing a write of records to maintain atomicity for writing a defined group of records to multiple tracks

Provided are a computer program product, system, and method for mirroring writes of records to maintain atomicity for writing a defined group of records to multiple tracks. Sub-writes are received from a host system to write records in defined groups in a write, wherein at least one of the sub-writes includes records to write for at least one of the defined groups that spans multiple tracks. For each of the sub-writes including records for at least one of the defined groups that spans multiple tracks, the tracks in the sub-write with to the secondary storage system to cause the secondary storage system to apply the tracks upon receiving all the tracks for the sub-write. Complete is returned for the sub-write to the host system upon receiving confirmation from the secondary storage system that all the tracks for the sub-write have been applied to the secondary storage system.

Mirroring writes of records to maintain atomicity for writing a defined group of records to multiple tracks

The invention provides an apparatus, system, and method for cluster-wide peer recovery in the event of a computer failure. A failure of a first computer is detected and a recovery module is registered as the first computer. In one embodiment, the recovery module is a peer computer. The recovery module retrieves a privately held undo log data through the authorized assumption of the failure identity associated with the failed first computer, backs out in-flight transaction updates of the first computer, and frees up data resources locked by the first computer.

Apparatus, system, and method for transactional peer recovery in a data sharing clustering computer system

Growing community of inventors

San Martin, CA, United States of America

Ronald E Bretschneider