Gravic, Inc.

Intel Corporation

Rpx Corporation

John R Hoffmann

Bruce D Holenstein

Paul J Holenstein

Wilbur H Highleyman

Dylan R Holenstein

Gary E Strickler

Jonathan M Ziegler

Methods are provided for preventing constraint violations at a target database by a database replication system. The database replication system replicates database transactions from a change log of a source database to the target database. Each source database and each target database have one or more tables, and at least one of the target database tables has one or more constraints. The database replication system replicates database transactions from a change log for the source database to the target database. Each of the source database tables has an associated change log disk cache. For each database transaction, a database change is detected to the at least one of the source database tables that potentially causes a constraint violation at the target database as a result of one or more of the target database tables having one or more constraints. An application call, or a programmatic call by a database management system, is then made when the database change is detected to flush to a change log disk contents of the associated change log disk cache for the at least one of the source database tables which has a change log disk cache that may flush to the change log disk in a misordered manner compared to change log disk cache flushes from at least one other source database table. The application call, or the programmatic call by the database management system, prevents the potential constraint violation at the target database by the database replication system.

Methods for preventing constraint violations at a target database by a database replication system where an application call, or a programmatic call by a database management system, forces a flush of a change log

Commit latency is reduced for pending transactions. The transactions are replicated from a source database to a target database by a replication engine. The pending transaction is incomplete in either the source or the target database. A transaction pattern table is provided that includes a plurality of records, each record being defined by a transaction pattern identifier. Each record includes a transaction pattern and a metric of occurrence of the transaction pattern. A comparison is made of the series of one or more transaction steps or operations of a pending transaction to the transaction patterns in the transaction pattern table, and a subset of transaction patterns in the transaction pattern table are identified that are potential candidates for matching the pending transaction. The pending transaction is categorized using the metric of occurrence of the transaction pattern for each transaction pattern in the subset of transaction patterns. The categorizing predicts when all of the series of one or more transaction steps or operations which have completed in the source database have been replicated. The metric of occurrence of the transaction pattern for each record of the transaction pattern table is (i) a count that is continuously updated based on completed transactions that match the transaction pattern, or (ii) a prepopulated probability of occurrence of the transaction pattern, or (iii) derived from historical data of the transaction pattern. The pending transaction is then processed using the categorization of the pending transaction.

Apparatus, method, and computer program product for reducing commit latency of pending transactions

An automated method and apparatus are provided for comparing a first database and a second database that are intended to be kept in synchronization with each other. Applications process changes that are posted to the first database and the second database. Changes made to the first database are applied to the second database, and vice-versa. Changes made to the first database resulting from changes posted to the first database are collected into a first change log. Changes made to the second database resulting from changes posted to the second database are collected into a second change log. Representations of the changes that correspond to the same respective changes in the change logs are compared with each other. Corrective action is performed when the comparison indicates that the corresponding representations of the changes in the first change log and the second change log do not match.

Method and apparatus for continuously comparing two databases which are actively being kept synchronized

Hung transactions which are completed within a transaction processing system are resolved. The transactions are replicated from a source database to a target database by a replication engine. The hung transaction is incomplete in either the source or the target database. A transaction pattern table is provided that includes a plurality of records, each record being defined by a transaction pattern identifier. Each record includes a transaction pattern and a metric of occurrence of the transaction pattern. A comparison is made of the series of one or more transaction steps or operations of a hung transaction to the transaction patterns in the transaction pattern table, and a subset of transaction patterns in the transaction pattern table are identified that are potential candidates for matching the hung transaction. The hung transaction is categorized using the metric of occurrence of the transaction pattern for each transaction pattern in the subset of transaction patterns. The categorizing predicts when all of the series of one or more transaction steps or operations which have completed in the source database have been replicated. The metric of occurrence of the transaction pattern for each record of the transaction pattern table is (i) a count that is continuously updated based on completed transactions that match the transaction pattern, or (ii) a prepopulated probability of occurrence of the transaction pattern, or (iii) derived from historical data of the transaction pattern. The hung transaction is then processed using the categorization of the hung transaction.

Apparatus, method, and computer program product for resolving hung transactions

A method and apparatus are provided for avoiding target database constraint violations in a database replication system. The database replication system includes a source database and a target database, wherein target database transactions automatically abort due to constraint violations. Each database has one or more tables and at least one of the target database tables has one or more constraints. The database replication system replicates source database transactions. For each target database transaction corresponding to a replicated source database transaction, changes made at the source database are replicated to the target database. When it is detected that a replicated change causes a constraint violation as a result of one or more of the target database tables having one or more constraints, the applying of the database change to the target database of the database change is deferred and the database change is posted to an electronic list. Replicated changes that do not cause a constraint violation are immediately applied to the target database. This process is repeated for all of the changes within the replicated source database transaction. The database changes on the electronic list are iterated through by applying the database changes on the electronic list that no longer cause a constraint violation to the target database and removing the applied database changes from the electronic list, and by deferring again the applying of the database changes on the electronic list that still cause a constraint violation and maintaining the deferred database changes on the electronic list. A commit is performed for the target database transaction corresponding to the replicated source database transaction when a commit operation is detected in the replicated source database transaction, and when no database changes remain on the electronic list for the replicated source database transaction. The deferring process avoids the target database transactions automatically aborting due to constraint violations.

Method and apparatus for resolving target database constraint violations in a database replication system where target database transactions are automatically aborted due to constraints violations

An automated method and apparatus are provided for comparing a source database and a target database that are intended to be kept in synchronization with each other. An application processes database changes that are posted to the source database and replicated to the target database. The source database and the target database are made up of blocks of data. Blocks of data from the source database having database changes are collected. These blocks of data are replicated to the target database. Blocks of data from the target database that correspond to the blocks of data from the source database having database changes are read. Representations of the blocks of data from the source database having database changes are compared to representations of blocks of data from the target database that correspond to the blocks of data from the source database having database changes. A corrective action is performed when the comparison indicates that the representations of the blocks of data from the source database having database changes do not match the representations of blocks of data from the target database that correspond to the blocks of data from the source database having database changes. A similar process is performed using 'before images' of blocks of data from the source database having database changes and then comparing representations of these “before images” to representations of “current images” of blocks of data from the target database that correspond to the before images of blocks of data from the source database having database changes.

Method and apparatus for using representations of blocks of data when continuously comparing two databases which are actively being kept synchronized

An automated method and system are provided for comparing a source database and a target database that are intended to be kept in synchronization with each other. An application processes transactions that are posted to the source database and replicated to the target database. Before images of changes made to records or rows of the source database resulting from transactions posted to the source database into a first change log are collected. The before images of changes made to records or rows of the source database are replicated to the target database. Current images of records or rows of the target database that correspond to the before images of changes made to records or rows of the source database are read. Representations of the before images of changes made to records or rows of the source database are compared with representations of the current images of records or rows of the target database that correspond to the before images of changes made to records or rows of the source database. A corrective action is performed when the comparison indicates that the representations of the before images of changes made to records or rows of the source database do not match the representations of the current images of records or rows of the target database that correspond to the before images of changes made to records or rows of the source database.

Method and system for using before images of replicated changes from a source database with current target database images read from the target database when continuously comparing two databases which are actively being kept synchronized

Methods and apparatus are provided for processing a transaction that has been aborted on a source system. Transactions are completed within a transaction processing system, and the transactions are replicated from a source database to a target database by a replication engine. Each transaction is a series of one or more transaction steps or operations. The aborted transaction on the source system was replicated and is a pre-committed transaction on the target database. Transaction pattern data in a transaction pattern table is used to predict when a pre-committed transaction is an erroneously committed transaction. The pre-committed transaction is reversed on the target database when the prediction indicates that the pre-committed transaction was erroneously committed.

Systems and methods to process an erroneously pre-committed transaction on a target database which has been aborted on a source database and replicated from the source database to the target database

An automated method and system are provided for comparing a source database and a target database that are intended to be kept in synchronization with each other. An application processes transactions that are posted to the source database and replicated to the target database. Before images of changes made to the source database resulting from transactions posted to the source database are collected into a first change log and then replicated to the target database. Before images of changes made to the target database resulting from replication are collected into a second change log. Representations of the before images of changes in the first change log are compared with representations of the before images of changes in the second change log that correspond to the same respective changes that were made to the source database and replicated to the target database. Corrective action is performed when the comparison indicates that the representations of the before images of changes in the first change log do not match the representations of the before images of changes in the second change log.

Method and system for using before images of changes for continuously comparing two databases which are actively being kept synchronized

Methods and apparatus are provided to fully process an incomplete transaction which is incomplete in either a source database or a target database. Transactions are completed within a transaction processing system that includes a non-static application. The non-static application has a transaction mix that varies over its life cycle. The transactions are replicated from the source database to the target database by a replication engine. Each transaction is a series of one or more transaction steps or operations. A plurality of transaction pattern tables are provided, and one of the transaction pattern tables are selected for use in the process depending upon a time of day or a transaction mix.

Systems and methods to fully process an initially incomplete replicated and committed transaction for a non-static application by using a plurality of transaction pattern tables

An automated method is provided for use when replacing a currently operating data replication engine in a first system with a new data replication engine in the first system in a bidirectional data replication environment. The currently operating data replication engine in the first system and the new data replication engine in the first system replicates first database transactions from an audit trail of a first database in the first system to a second database in a second system. The new data replication engine in the first system generating a list of active database transactions in the first system, and sends the list of active database transactions to the new data replication engine in the second system as a first token. The new data replication engine in the second system receives the first token, fetches a transaction event from an audit trail of second database, and replicates the fetched transaction event to the new data replication engine of the first system when the fetched transaction event does not match a transaction on the list in the first token. These steps are repeated during operation of the new data replication engine of the second system. The currently operating data replication engine in the first system is stopped from replicating first database transactions when all of the transactions on the list of active database transactions that were generated have been replicated to the second system.

Method for replacing a currently operating data replication engine in a bidirectional data replication environment without application downtime and while preserving target database consistency, and by using audit trail tokens that provide a list of active transactions

Methods and apparatus are provided to fully process an initially incomplete replicated and committed transaction. A transaction pattern table is provided that includes a plurality of records, each record being defined by a transaction pattern identifier. Each record includes a transaction pattern and a metric of occurrence of the transaction pattern. An incomplete transaction is automatically compared to the transaction patterns in the table and a subset of transaction patterns in the table are identified that are potential candidates for matching the incomplete transaction. The incomplete transaction is categorized using the metric of occurrence of the transaction pattern for each transaction pattern in the subset. The categorization predicts when all steps or operations of the transaction which have completed in the source database have been replicated, and the incomplete transaction is committed based on this prediction. Any additional steps or operations which have completed in the source database for the committed incomplete transaction are processed as a new transaction applied to the target database, thereby fully processing the incomplete transaction.

Systems and methods to fully process an initially incomplete replicated and committed transaction

Source database precommitted transactions are resolved in a target database of a database replication system when selected source database precommitted transactions are subsequently aborted in the source database.

Method and apparatus for resolving source database precommitted transactions that are replicated to a target database of a database replication system

An automated method is provided for replacing a currently operating data replication engine with a new data replication engine. The currently operating data replication engine replicates source database transactions including any new source database transactions committed in an audit trail of the source database before the currently operating data replication engine stops replicating. All new source database transactions that the currently operating data replication engine has not completed replicating are joined. The currently operating data replication engine is stopped when all of the source database transactions that were not joined have completed replicating. All joined transactions are then voted to be committed and are replicated. The new data replication engine replicates all new source database transactions started after the currently operating data replication engine stops.

Method for replacing a currently operating data replication engine with a new data replication engine without application downtime and while preserving target database consistency, and by joining the source database transactions

Target database constraint violations are potentially resolved in a database replication system by replacing, converting, or removing deferred database changes.

Method and apparatus for potentially resolving target database constraint violations in a database replication system by replacing, converting or removing deferred database changes

Growing community of inventors

Kennett Square, PA, United States of America

John R Hoffmann