As described earlier, a replication group is a collection of one or more objects, generally tables, that are logically related and that can or should be administered together. As a very general guideline, you can place all replicated objects associated with a given application into a single group. A common practice is to create all of an application’s objects under a single schema (i.e., Oracle user account); in such cases, it often makes sense to use a one-to-one correspondence between schema and replication groups. In fact, prior to Oracle 7.3, that was not only the assumption but also the requirement.
Oracle furnishes the following built-in package procedures to create and drop replication groups:
- DBMS_REPCAT. CREATE_MASTER_REPGROUP
Creates a replication group. The group is initially quiesced and contains no objects.
- DBMS_REPCAT. DROP_MASTER_REPGROUP
Drops a replication group.
- DBMS_REPCAT. COMMENT_ON_REPGROUP
Creates or replaces a comment on a replication group, visible in the DBA_REPGROUP data dictionary view.
Refer to Appendix A, for the complete API reference to these procedures.
The data dictionary view DBA_REPGROUP contains information about all replication groups at the current site.
The call to DBMS_REPCAT.CREATE_MASTER_REPGROUP automatically creates an entry in the job queue to process tasks in the repcatlog queue. This scheduled job calls DBMS_REPCAT.DO_DEFERRED_REPCAT_ADMIN every 10 minutes.
Oracle replication has many different ...