This procedure splits the workload of an existing Extract group into multiple Extract groups. It also provides instructions for including a data pump group (if applicable) and a Replicat group to propagate data that is captured by the new Extract group.

Steps are performed on the source and target systems.

1. Make certain the archived transaction logs are available in case the online logs recycle before you complete this procedure.

2. Choose a name for the new Extract group.

3. Decide whether or not to use a data pump.

4. On the source system, run GGSCI.

5. Create a parameter file for the new Extract group.

   EDIT PARAMS group
   

   Note:

   You can copy the original parameter file to use for this group, but make certain to change the Extract group name and any other relevant parameters that apply to this new group.

6. In the parameter file, include:

   • EXTRACT parameter that specifies the new group.

   • Appropriate database login parameters.

   • Other appropriate Extract parameters for your configuration.

   • EXTTRAIL parameter that points to a local trail (if you will be adding a data pump) or a RMTTRAIL parameter (if you are not adding a data pump).

   • RMTHOST parameter if this Extract will write directly to a remote trail.

   • TABLE statement(s) (and TABLEEXCLUDE, if appropriate) for the tables that are to be processed by the new group.

7. Save and close the file.

8. Edit the original Extract parameter file(s) to remove the TABLE statements for the tables that are being moved to the new group or, if using wildcards, add the TABLEEXCLUDE parameter to exclude them from the wildcard specification.

9. (Oracle) If you are using Extract in integrated mode, register the new Extract group with the source database.

   REGISTER EXTRACT group DATABASE [CONTAINER (container[, ...])]
   

10. Lock the tables that were moved to the new group, and record the timestamp for the point when the locks were applied. For Oracle tables, you can run the following script, which also releases the lock after it is finished.

    -- temp_lock.sql
    -- use this script to temporary lock a table in order to
    -- get a timestamp
    
    lock table &schema . &table_name in EXCLUSIVE mode;
    SELECT TO_CHAR(sysdate,'MM/DD/YYYY HH24:MI:SS') "Date" FROM dual;
    commit;
    

11. Unlock the table(s) if you did not use the script in the previous step.

12. Stop the old Extract group(s) and any existing data pumps.

    STOP EXTRACT group
    

13. Add the new Extract group and configure it to start at the timestamp that you recorded.

    ADD EXTRACT group, TRANLOG, BEGIN YYYY/MM/DD HH:MI:SS:CCCCCC
    

14. Add a trail for the new Extract group.

    ADD {EXTTRAIL | RMTTRAIL} trail, EXTRACT group
    

    Where:

    • EXTTRAIL creates a local trail. Use this option if you will be creating a data pump for use with the new Extract group. Specify the trail that is specified with EXTTRAIL in the parameter file. After creating the trail, go To Link a Local Data Pump to the New Extract Group.

    • RMTTRAIL creates a remote trail. Use this option if a data pump will not be used. Specify the trail that is specified with RMTTRAIL in the parameter file. After creating the trail, go To Link a Remote Replicat to the New Data Pump

    You can specify a relative or full path name. Examples:

    ADD RMTTRAIL dirdat/rt, EXTRACT primary
    ADD EXTTRAIL c:\ogg\dirdat\lt, EXTRACT primary

To Link a Local Data Pump to the New Extract Group

1. On the source system, add the data-pump Extract group using the EXTTRAIL trail as the data source.

   ADD EXTRACT pump, EXTTRAILSOURCE trail
   

   For example:

   ADD EXTRACT pump2, EXTTRAILSOURCE dirdat\lt
   

2. Create a parameter file for the data pump.

   EDIT PARAMS pump
   

3. In the parameter file, include the appropriate Extract parameters for your configuration, plus:

   • RMTHOST parameter to point to the target system.

   • RMTTRAIL parameter to point to a new remote trail (to be specified later).

   • TABLE parameter(s) for the tables that are to be processed by this data pump.

4. In GGSCI on the source system, add a remote trail for the data-pump. Use the trail name that you specified with RMTTRAIL in the parameter file.

   ADD RMTTRAIL trail, EXTRACT pump
   

   For example:

   ADD RMTTRAIL dirdat/rt, EXTRACT pump2
   

5. Follow the steps given below.

To Link a Remote Replicat to the New Data Pump