Data movement service configuration

The behavior and scheduling of each data movement services component can be configured to meet the different needs of a development, test, and production environment. Changing the configuration of one component can have a direct impact on the behavior of other components that are dependent on that component.

In general, there are two dependencies:
Because the dependent components need to operate on different schedules than the component they are dependent on, an invocation does not necessarily result in execution. Instead, each dependent component checks its schedule upon invocation and returns control to the calling component if it is not yet time to perform any tasks. In the above example, ETL and Target Llife Cycle components might only be executed twice if the schedule for both components restricts them from being invoked more than once every 5 time units.
ETL and Target Life Cycle component
invocation and execution
The ETL component (and Target Life Cycle component) is invoked and executed at T2 (T3 respectively). The next invocation occurs at approximately T6. Because fewer than 5 time units have passed since they last executed, control is immediately returned to the Apply component. Subsequent invocations at approximately T8 (T9 respectively) result in execution because more than 5 time units have passed. Each component is implemented by one or more component instances. You can configure each one of the instances separately to allow for more granular control.
Note: If changes are made, they take effect immediately, unless noted otherwise.

You can modify the default configuration for the Capture and Apply components by changing the appropriate control tables or by overriding them using command-line parameters in the start scripts. You can configure the ETL and life cycle enforcement components by updating one of the control tables.

You perform the following steps to customize data movement service components to meet the requirements of development, test, and production environments.

Configuring the (Source) Capture component instances

A Capture component instance is equivalent to a DB2® Capture replication utility. By default, this utility is configured to continuously capture changes to the source tables and record the changes in internal work tables. In general, you do not need to change the default configuration for the Capture component instances.

Configuring the Source Life Cycle component

Multiple life cycle component instances are being used in each source database (State and Runtime databases). Each instance, which is implemented by a trigger, enforces the retention policies as defined in the control table WBIRMADM.RMPRUNECTRL located on the source database for that data movement service. Life cycle retention policies are specified on a per table basis. Thus, one row in WBIRMADM.RMPRUNECTRL corresponds to one table requiring pruning.
Table 2.

Configuring the (Target) APPLY Component

An instance of an Apply component is a DB2 Apply replication utility. Changes that have been captured by Capture utilities are continuously applied to staging tables in the target database by default. The default utility configuration parameters should be sufficient for most environments and should not be changed.

Configuring the ETL Component

ETL components have been implemented in WebSphere® Business Monitor as database stored procedures. These stored procedures always reside on the target database for any given data movement service. Therefore, all ETL stored procedures assigned to the State to Runtime data movement service are located in the Runtime database, and ETL stored procedures assigned to the Runtime to Historical data movement service reside in the Historical database.

Configuring the Target Life Cycle component.

ETL work tables continuously grow as long as new or updated data is applied by Apply component instances. One Target life Cycle component instance, implemented by a stored procedure, is assigned to one work table in each target (Runtime and Historical) database. Each instance enforces the internal retention policies as defined in control table WBIRMADM.RMPRUNECTRL. As in the source tables, life cycle retention policies for ETL work tables are specified on a per table basis. Thus, one row in WBIRMADM.RMPRUNECTRL corresponds to one table that requires pruning.

Summary of the configuration parameters of data movement services

The following table summarizes the most commonly used parameters provided for each of the data movement services components. For more information about configuration parameters, refer to DB2 Replication guide.
Component Parameter name Default values Valid values Parameter location
Capture autoprune Y    
Capture prune_interval (seconds) 300    
Source Life Cycle PRUNE_ENABLED 1

0 - Disabled

1 - Enabled

Data movement service Source DB: WBIRMADM.RMPRUNECTRL

Source Life Cycle RETENTION_IN_MINUTES

0 - State to Runtime

1440 - Runtime to Historical

0 to DB2 limit for BIGINT

Data movement service Source DB: WBIRMADM.RMPRUNECTRL

Source Life Cycle PRUNE_INTERVAL (minutes) 5 0 to DB2 limit for BIGINT

Data movement service Source DB: WBIRMADM.RMPRUNECTRL

ETL ETLSCHEDMETHOD 1

0 - Flexible scheduling

1 - Strict interval scheduling

Other - Disables ETL

Data movement service Target DB: WBIRMADM.RMCONTROL

ETL ETL_0_MINUTES

5 - State to Runtime

1440 - Runtime to Historical

0 to DB2 limit for INTEGER

Data movement service Target DB: WBIRMADM.RMCONTROL

ETL LOGLEVEL 0

0 - For normal logging

1 - For trace logging

Data movement service Target DB: WBIRMADM.RMCONTROL

ETL COMMITINTERVAL (number of records.) 1000

0 - Disable commits until the end

1 - Commit every record.

n - DB2 Limit for BIGINT

Data movement service Target DB: WBIRMADM.RMCONTROL

Target Life Cycle PRUNE_ENABLED 1

0 - Disabled

1 - Enabled

Data movement service Target DB: WBIRMADM.RMPRUNECTRL

Target Life Cycle RETENTION_IN_MINUTES 0 0 to DB2 limit for BIGINT

Data movement service Target DB: WBIRMADM.RMPRUNECTRL

Target Life Cycle PRUNE_INTERVAL (minutes) 1440 0 to DB2 limit for BIGINT

Data movement service Target DB: WBIRMADM.RMPRUNECTRL

Note: IBM® reserves the right to make changes to the database tables and columns referenced above. As such, some tables and columns may be altered, removed, or added from release to release. Any reliance on the content or structure referenced in the information from release to release is done at the consumer's own risk. IBM will document any such changes as they occur.
Related tasks
Deploying data movement services
Configuring data movement services options
Finalizing data movement services setup
Consolidating start and stop scripts

Copyright IBM Corporation 2005. All Rights Reserved.