Siebel Remote and Replication Manager Administration Guide > Implementing Replication Manager > Extracting the Regional Database >

Extracting Regional Databases in Parallel

Parallel Database Extract is a version of the Database Extract server component that starts extractions in parallel. The Parallel Database Extract server component improves performance when Replication Manager extracts a large database for one or more regional nodes. It is an interactive component, using server infrastructure features to do data extractions in parallel for a large regional node. It uses multithreaded components to do this work more quickly than a single thread operation. Parallel Database Extract requires more hardware, but it can significantly reduce the time that Replication Manager requires to do a database extract.

To extract a regional database by using the Parallel Database Extract option, you can use the administrative interface or the command line interface.

Extracting Regional Databases in Parallel by Using the Administrative Interface

You can extract regional databases in parallel by using the administrative interface.

To extract regional databases in parallel by using the administrative interface

1. Make sure advanced parameters are visible.

   For more information, see Making Advanced Server Parameters Visible.

2. Adjust the shared pool size.

   For more information, see Adjusting the Shared Pool Size for Parallel Database Extract.

3. Locate the record for the Parallel Database Extract server component.

   For more information, see Locating the Record for a Server Component.

4. Use Columns Displayed to expose the Max Tasks field of the Components list.
5. In the bottom applet, click the Parameters tab, click Advanced, and then query the Parameter field for the following value:

   Maximum Number of Work Queue Threads

   The Max Tasks field for the Parallel Database Extract server component displays the maximum number of worker threads for each work queue. The Maximum Number of Work Queue Threads parameter determines the degree of parallelism.

6. Complete the procedure for using the administrative interface that is described in Extracting the Regional Database by Using the Single Thread Option, with the following modifications:
   • In the Component/Job field, choose Parallel Database Extract.
   • To specify more than one regional node, you can create a text file and then enter the name of each regional node on a single line in this file. When you specify the Value for the Client Name parameter, you enter the path and file name, preceded by the at sign (@). For example, a Windows client might specify the following value as the Value for the Client Name:

   @D:\workdir\regnodes.txt

Extracting Regional Databases in Parallel by Using the Command Line Interface

You can extract regional databases in parallel by using the command line interface.

To extract regional databases in parallel by using the command line interface

• From the Server Manager command line, enter the following command, all on one line:

  start task for comp pdbxtract server server_name with client=regional_node_name, initmethod=ddl

  For more information, see Opening the Command Line Interface for Server Manager.

  To specify more than one regional node, you can use a command to separate each regional node name. For example, if RN1, RN2, and RN3 are the names of three regional nodes, then you enter the following command:

  start task for comp pdbxtract server server_name with client=RN1,RN2,RN3, initmethod=ddl

Adjusting the Shared Pool Size for Parallel Database Extract

To meet the cache requirements for a parallel database extract, the init.ora file contains a parameter named SHARED_POOL_SIZE that you can adjust. The number of temporary tables that you use depends on the size of the shared pool that the Siebel Server can access. If the size of the shared pool is less than 300 MB, it is recommended that you use one temporary table and run one instance of the Database Extract component. If the size of the shared pool is greater than 600 MB, then using one temporary table for each instance of the Database Extract component might increase throughput.

Table 26 describes some examples for setting the shared pool size.

For more about:

• Sizing the shared pool, see Siebel Performance Tuning Guide
• The init.ora file and the SHARED_POOL_SIZE parameter, see the Siebel Installation Guide for the operating system you are using