Metabases are repositories of database tables and objects that define, enable, and manage data distribution replication projects. Metabases are tied to a project and each project has a unique metabase. Metabases contain replication backlogs and metadata about what tables are enabled for capture on the source. Db2 for IBM i, Db2 for LUW, Oracle, and data connections used for a replication project must be associated with a metabase.
You create metabases in Data Integration to be saved on the database server accessed by the JDBC data connection. Metabases are stored in the database on the source system. Each time replication project changes are applied, information is written to the metabase.
The metabase includes:
- Log reader capture/status information
- Sync request copy status information. It should be highlighted that a metabase can only be used with one project. This is noted for iSeries, but the same condition applies to any DBMS connection.
- One or more indexes, views, trigger tables,and data replication backlog tables. The backlog tables contain changed data updates by column that are linked together by row, table, and transaction ID, queued (waiting) to be replicated.
- The status and number of tables that are enabled for capture.
- Information about the replication project, stored in memory.