Recovery method for incremental data synchronization fault in isomerous database synchronizing system

Abstract

The invention discloses a recovery method for an incremental data synchronization fault in an isomerous database synchronizing system. The isomerous database synchronizing system comprises at least two databases. The recovery method comprises the following steps that (1) a corresponding special table is established in each goal database, wherein the special table of each goal database is used for recording the synchronization position of a source database which is synchronous with the goal database; (2) a transaction execution mode of each goal database is set to be a manually submitting mode, after synchronization of incremental data is completed, the synchronization position of the corresponding special table is updated, and then the incremental data transaction is manually submit; (3) if a fault is caused when incremental data synchronization is conducted from each source database to the corresponding goal database, the synchronization position of the source database is read from the special table of the corresponding goal database, and then incremental data synchronization continues with the synchronization position as a start point from the source database to the corresponding goal database. According to the recovery method for the incremental data synchronization fault in the isomerous database synchronizing system, the performance of the databases is hardly influenced, and quick recovery of the synchronization fault is conducted through the atomicity of transactions of the databases.

Description

technical field [0001] The invention relates to the technical field of database synchronization, in particular to a recovery method for incremental data synchronization faults in a heterogeneous database synchronization system. Background technique [0002] The database is a basic component in the IT architecture and an effective facility for storing and managing data. Now almost all applications are inseparable from the support of the database. Due to the uneven distribution of massive data and data access areas, a single database can no longer carry all the access. Generally, a database needs to be deployed in multiple different places to provide services. Each database is data complete and reliable to the outside world. The service is complete, but in fact the data between each two databases is inconsistent, so incremental data synchronization between these databases in different places is required to ensure the data consistency and service integrity of all databases. ...

AI Technical Summary

Problems solved by technology

[0006] The key to recovering the incremental data synchronization that has been interrupted is to record the synchronization location information. Using files to record synchronization location information is a commonly used method, but since the file operation and the database operation are not the same atomic operation, the write file operation and the write database operation There is a very small time interval between them. In some extreme cases (such as a power failure when the file operation has just been written and the database operation is not performed), it will still damage the consistency of the database. Therefore, if you use a file to record the synchronization position, you must Design a more complex method to ensure the consistency of incremental data and synchronous position

Images