|Replication Operation Support|
|Schema Replicated||Yes, with limitations|
|ddlscan Supported||Yes, supported for mixed Oracle/MySQL, and data warehouse targets|
To extract information from an Oracle database two options are available,
the CDC process (for more information, see
Section 5.5, “Deploying Oracle Replication using CDC”), and a system that reads
information directly from the Oracle redo log. This includes another process
that runs alongside the Replicator and extracts events from the Oracle Redo
Log. The functionality listed here is only available in the release named
Contact your support if you don't have access to that release package.
Replication with the Oracle Redo Reader operates as follows (see Figure 5.1, “Internals: Oracle Extraction with Oracle Redo Reader” for a graphical representation):
A separate Oracle Redo Reader (orarr) process accesses and parses the content from the Oracle redo, archive, and supplemental logs, generating the individual transaction information into a format called the PLOG. The information in the PLOG contains each individual transaction, including limited DDL statements, in a format that can be processed by the Tungsten Replicator. This component is configured and installed automatically during the standard Tungsten Replicator deployment process.
The Tungsten Replicator reads the information from the PLOG and converts it into the Transaction History Log (THL) used by the replicator to store events and transactions. Once the transactions have been converted into the internal THL format, the information can be used by any downstream replicator to apply the data into another Oracle instance, or heterogeneous target such as MySQL or data warehouses.
The Oracle Redo Reader process maintains a repository of it's own internal
configuration, and progress, within a tracking schema in the source
database. By default, this tracking schema will be called
ORARRUSER and is created automatically during
The Oracle Redo Reader process maintains a repository of it's own internal configuration, and progress, within a tracking schema in the source database. By default, this tracking schema will be called ORARRUSER and is created automatically during install.
The Tungsten Replicator maintains a separate tracking schema with a small number of tables used to track progress of replication. The extractor writes to this tracking schema, the Applier reads from this table during startup to establish the correct position to begin replication. The account needs to be pre-created before installation and is details in Section B.5, “Oracle Database Setup”.
Tungsten Replicator with Redo Log processing is supported within the following environments:
Oracle 9.2 (Experimental), 10g, 11g, 12c
Enterprise Edition (EE), Standard Edition (SE), Standard One (SE1) and Express Edition (XE)
Oracle RAC (Experimental)
In addition, the following requirements and limitations are in effect: