CDC Guide for Linux, UNIX, and Windows

10.2
- 10.2.2
- 10.2.1
- 10.0

Back Next

Configuration in an Oracle RAC Environment

PowerExchange Oracle CDC with LogMiner can process change data for database instances in a real application cluster (RAC) environment.

For CDC to work properly, you might need to apply certain Oracle patches. In particular, Oracle bug 6596564 can cause data loss or data integrity problems in an Oracle RAC environment where CDC runs. If you use Oracle 11.1.0.7 on Windows, install Patch 24 before starting change capture. An Oracle 11.1.0.7 patch is not available for Linux and UNIX. If you use Oracle 11.2.0.1 or later, you do not need to apply a patch because the 6596564 bug fix is part of the base version.

Also complete the following configuration tasks:

Ensure that the Oracle instance from which you run PowerExchange CDC can access the Oracle archived redo logs for all Oracle instances in the RAC from which change data is captured.

In the init.ora file for each of these Oracle instances, define the LOG_ARCHIVE_DEST_1 parameter to point to the directory in which Oracle creates the archived logs.

PowerExchange uses Oracle LogMiner to read change data from the archived logs. If you use an archived log destination other than the LOG_ARCHIVE_DEST_1 path and LogMiner processing lags behind, problems might occur. In this situation, LogMiner starts reading change data from the archived logs in the LOG_ARCHIVE_DEST_1 directory. If these archived logs are inaccessible from the machine with the Oracle instance to which you are connected, the LogMiner session might fail.

Ensure that all of the Oracle instances in the RAC that participate in CDC have access to the Oracle online redo logs. Usually, these redo logs reside on shared storage.

The tasks for configuring access to archived redo logs vary by operating system, as follows:

On Windows, set up an Oracle flash recovery area on the shared file system that contains all of the table data for the RAC. For each Oracle instance in the RAC, set the LOG_ARCHIVE_DEST_1 parameter to point to that recovery area.

On Linux and UNIX, use any of the following methods:

Set up an Oracle flash recovery area in the same manner as for Windows.

Store all archived redo logs on shared storage.

Set up Network File System (NFS) access to the archive logs.

If you use shared storage or NFS access, the Oracle instance from which you run CDC must access the archived logs of the other RAC member instances. This access uses the mount points that match the archive log directories defined for those member instances. For example, ORA1 is the Oracle instance that runs CDC, and ORA2 is another Oracle instance in a RAC. ORA2 has a LOG_ARCHIVE_DEST_1 parameter that points to the following archive log directory: