CDC Provider Replication Server properties list

The following tables list the required and optional properties of the component.


Required properties

Property

Description

RepServer Name

Specify the Replication Server to use to capture data changes. After Replication is created, this property cannot be modified. If you want to modify it, right-click the component and select Drop Replication.

Rep Database

Object name generated during replication creation. This property is read-only.

Source Table Options

Set the replication definition options:

  • Interface – identify the method or driver you want to use to connect to the primary data source.

  • Host Name – identify the primary data source. The options in the Host Name list depend on the selected interface.

  • User and Password – identify an authorized database user, and protect the database against unauthorized access.If the source database is ASE, the database user requires “sa” or “dbo” permission or “replication_role.”

  • Database – identify the database to use as data source.If you select this option, you must also select an appropriate interface, and in some cases, specify an appropriate user ID and password.

  • Schema – identify the schema or owner you want to use as data source. The objects that appear are restricted accordingly and new tables are created in that schema.

  • Table – identify the source table.

  • Replicate – identify the column to be replicated.

    NoteSince ETL does not support Large Object (LOB) data processing, a column with LOB type is not replicated.

  • Key – identify the columns with keys. Replication Server and ETL use the key to identify each row of the source table.

    NoteReplication Server does not allow using column with LOB type as key. The key values should be unique for each row and the key columns must be replicated.

  • Size – identify the column size. You can change it to an appropriate value.

If the source database is Oracle, enter Replication Agent details:

  • RepAgent Host – Replication Agent name, which must be identical to the configuration value of ‘rs_source_ds’ in Replication Agent.

  • RepAgent Database – Replication Agent database name, which must be identical to the configuration value of ‘rs_source_db’ in Replication Agent.

  • RepAgent Username and Password – user name and password to connect to Replication Agent.

After Replication is created, this property cannot be modified. If you want to modify it, right-click the component and select Drop Replication.

Port Options

Specify port type of each OUT-port:

  • Port Name – specify the OUT-port name.

  • Function – specify the data changes to be sent from this port:

    • Insert – sends insert changes.

    • Delete – sends delete changes.

    • Update – sends update changes.

    • Upsert – sends insert and update changes.

    • All – sends all data changes.

Capture Mode

Specify whether to send all received data changes or only the last received data changes on each row, to the next component. By default, this property is set to “Last,” which is appropriate for most incremental loading.

Replication

Indicates whether replication is created or dropped. The value for this property is generated by ETL and is read-only.


Optional properties

Property

Description

Rep Server User and Password

Specify an authorized Replication Server user name and password.

Stage Mode

Specify whether data changes should be staged in the memory or in an IQ temporary table. If the capture mode property is set to Last, provide a value for the stage mode property.

NoteFor a 32-bit grid engine, it is recommended to set stage mode to IQ.

IQ for Stage Mode Options

Specify configuration for staging IQ. This property is enabled only when capture mode is Last, and stage mode is IQ.

Timeout seconds for no data

Specify the wait time, in seconds, for the component before it stops waiting to receive data changes from Replication Server.

Read Block Size

Determine the number of records retrieved by the component in a single step.

Auto Initial Load

Specify whether you want the initial load to take place automatically. Before transferring source table data changes, ETL must perform an initial load to synchronize the source and destination table data, and create replication objects to make Replication Server start capturing data changes.

NoteSince Replication Server does not support materialization for Oracle, this property does not work if the source database is Oracle.

While executing or simulating a project:

  • If this property is selected and Replication is not created, then ETL creates the replication definition, connection, and function strings, and uses the Atomic materialization method to create subscription. This materialization method locks the source table until all source table records are captured by Replication Server as insert operations in a transaction. After this, simulation or execution performs the initial load.

  • If this property is not selected and Replication is not created, you see an error message.

  • If this property is selected and execution or simulation fails, right-click the component and select Drop Replication. Try to execute or simulate the project again.

If you right-click the component and select Create Replication, while this property is selected, initial load does not take place.

NoteWhile simulating a project, if this option is selected and you click Reset and Start to restart simulation, CDC Provider Sybase Replication Server component does not receive the source table initial data a second time. Replication Server pushes initial data to ETL only once.

Output Old Value

Select to output old column values of each data change. After Replication is created, this property cannot be modified. If you want to modify it, right-click the component and select Drop Replication.

Transactional

All work performed by the CDC Provider Sybase Replication Server component, including pre-SQL and post-SQL, is done in a single database transaction that is committed when the project finishes normally. Select this option to roll back the transaction, if this component encounters an error. See “Job components” and “Enabling transactionality for projects and jobs” for information on the “Propagate Rollback” property.

Save Interval

Specify the amount of time, in minutes, for which Replication Server should save the data changes after it is received by ETL. By default, Replication Server discards each data change after ETL receives it. Set this property to have Replication Server save data changes for a while after ETL receives them.

It is necessary to save data changes, since a project or job failure can cause the destination table changes to roll back, if project level transactionality is enabled. In this case, all source table data changes received are not applied to destination tables. In subsequent executions, the data changes that are not processed successfully in the previous execution, that still exist in the Replication Server, are sent to the CDC Replication Server Provider component, along with any new changes. This ensures that the destination tables are updated correctly.If the save interval is too short, some data changes from running a failed project are not sent to the destination components, causing the destination tables to be out of sync with the source table data.If the save interval is too long, duplicate data changes are sent to the component, leading to decreased performance, and increased resource usage.