With the Server Manager, you can perform continuous data replication. Continuous replication delivers source table changes to the target with very low latency.
Before you configure continuous replication, verify that Data Replication supports change data capture from online logs for the source database. Also, from the Data Replication Console, define the tasks that you want to execute continuously.
Load or create a configuration.
Switch to Edit mode.
On the
Extract Range
tab, select one of the following options if you want the Extractor to capture data from online logs:
For Microsoft SQL Server sources, select
Read from online transaction logs
.
For Oracle sources, select
Read from online redo logs
.
Click the
Runtime Settings
tab >
General
view.
In the
Continuous replication latency
field, enter the latency, in seconds, of change capture processing during continuous data replication.
This value determines the duration of an Extractor microcycle. During a microcycle, the Extractor captures changes from database logs, creates an intermediate file, writes changes to the intermediate file, and then sleeps. When the latency period ends, the Extractor begins another microcycle.
When setting the latency period, consider how it affects replication performance:
If the latency period is greater than the time it takes the Extractor to read changes from the database logs and write them to an intermediate file, the Extractor sleeps until the end of the latency period. To minimize inactive Extractor sleep periods, set the latency to a low value.
If the latency period is equal to or less than the time it takes the Extractor to read changes from the database logs and write them to an intermediate file, the Extractor continues processing and does not sleep. The Extractor begins a new microcycle as soon as it closes the intermediate file. However, if the latency is too low, performance of a long-running Extractor task might be degraded because of high disk memory or core loads. In this case, try increasing the latency value in sub-second increments such as 0.1 seconds.
Default is 0.333 seconds, which is usually acceptable to process all changes in a microcycle.
After the Extractor creates the intermediate file, the Applier can start reading changes and applying them to the target. Therefore, the Applier might also run during a microcycle. For large transactions with many changes, the Applier waits until the Extractor writes all of the changes to the intermediate file, which might take several microcycles, and then begins apply processing.
Save the configuration.
On the
Server Manager
tab >
Schedules
view, create a schedule with the tasks you defined for continuous replication.
Do not add subtasks that have dependencies on other tasks.
To create a schedule, use one of the following methods:
Click
New
to define a schedule in the
New
dialog box. In the
Run
field, you must select
Continuous
. Also define a run period under
Schedule Options
. In the
End date
field, you can specify a date with a year up to 2037.
Click
Wizard
to start the
Schedule wizard
. In Step 1, you must select
Continuous
as the execution scheme. In Step 2, define a run period under