International Business Machines Corporation

Phillip Koza

James D Spyker

Victor L Szabo

Mohamed M Ahmed

Laura A Cuell

Nirmal Kumar

Log reader and parser sharing determination in a change data capture environment includes obtaining log record identifiers for a group of data consumers of transaction log file data in transaction log files, identifying a respective restart log position and current log position for each data consumer of the group of data consumers, determining transaction log file distances between the restart log positions and the current log positions of each of one or more pairs of data consumers, calculating resource costs for data consumers of each pair to share a log reader and log parser, determining whether to share a log reader and log parser between any two or more data consumers of the group, and performing processing based on the determining whether to share.

Log reader and parser sharing determination in a change data capture environment

A computer-implemented method includes detecting a source update applied to a source table, where the source table includes a plurality of columns, and where the source update modifies a first subset of columns of the plurality of columns. A second subset of columns of the plurality of columns is determined, by a computer processor. The second subset of columns has more columns than the first subset of columns and fewer columns than the plurality of columns. Determining the second subset of columns includes determining which of the plurality of columns were updated most recently. The second subset of columns is included in a set clause of an update statement for updating a target table, where the target table is or includes a replica of the source table. The update statement is applied to the target table to achieve at least a minimum batching percentage within a plurality of update statements.

Replicating database updates with batching

According to one embodiment of the present invention, during database replication from a source database to a target database, operations applied to the source database are transferred as target operations and applied to the target database. A request from a requestor to idle a specified database object is received during the database replication. In response to receiving the request, transference of the target operations for the specified database object from the source database to the target database is terminated and target operations for the specified database object on the target database are monitored. In response to no target operations being applied to the specified database object on the target database, replication of the specified database object is suspended and the requestor is notified when the idling of the specified database object has completed.

Idling individually specified objects during data replication

A system and method manages resources within a process by identifying a first resource threshold associated with available resources. The method receives a resource request from a resource requesting participant requiring the available resources. The method determines that an estimated resource usage, associated with the resource request, approaches the first resource threshold. The method negotiates with at least one resource consuming participant to voluntarily release resources until the estimated resource usage approaches a second resource threshold associated with the available resources within the process.

Resource management via iterative negotiation

According to one embodiment of the present invention, a system reduces re-reading of database logs by persisting long running transaction data. The system processes the database log records of the database transaction log; generates a snapshot of the database log records at periodic intervals, wherein each snapshot includes database log records for pending transactions; and in response to an interruption in processing of the database log records, utilizes a snapshot to restore database log records for the pending transactions and resume processing of the database transaction log from a position succeeding the database log records of the selected snapshot.

Growing community of inventors

San Jose, CA, United States of America

Phillip Koza