DBReplicator API Reference


How DBReplicator Works

2.2. How DBReplicator Works

DBReplicator is a trigger-based database replication engine. It takes each insert/update/delete operation and creates a shadow record in a shadow table for that operation. These shadow records are numbered, and each subscriber maintains a bookmark for that table. Records are deleted from the shadow tables when they have been synced to all subscribers and publishers. Using this model, only changes are sent during normal sync sessions.

Being trigger-based, it works rather quickly, and has more options for resolving conflicts involving foreign keys, but it won't capture DDL (Data Definition Language) events such as changes to tables/views/sequences. There is a plan for capturing DDL changes by storing hasmaps of database metadata and comparing them with existing metadata, but this is not yet implemented. At present, DDL changes must be distributed manually.


Powered by Drupal - Theme by Danger4k