Top FAQs for Data Replication

Use these FAQs to learn more about data replication tasks including extracting and replicating data from Oracle Fusion Cloud Applications, uploading data to or downloading data from object storage, and loading data to the target database.

What can I do if a data replication job runs for a long time?

If a job runs for a long time, try the following:

• If the replicated view object (VO) isn't an extract VO (that is, the VO name doesn't end with ExtractPVO), then use the Data Replication editor to exclude unnecessary LastUpdateDate columns from the new data identifier of the VO.
• If the long running view object has more than one LastUpdateDate column selected for the new data identifier or incremental filter:
  • Select the LastUpdateDate option for the primary entity of the VO.
  • Deselect the LastUpdateDate option for columns from supplementary entities (non-functional).
• If you can't deselect the option for the new data identifier, then follow these steps:
  1. Cancel the job.
  2. Drop the Staging table TMP$.
  3. Navigate to the main menu, then click Data, then click Connections.
  4. Click Target Connection, select Inspect, then click the Tables tab.
  5. Select the table, then select the Reset Updated Time, then select Reload all data.

What can I do to improve the performance of the data replication job?

To improve performance, try the following:

• Only replicate with extract data stores (that is, View Objects (VOs) with "ExtractPVO" in the VO name.
• If the replicated VO is not an extract VO (that is, the VO name doesn't end with "ExtractPVO"), then use the Data Replication editor to exclude any unnecessary LastUpdateDate columns from the new data identifier of the VO.
• Make sure that the Public View Object (PVO) load type isn't set to FULL mode unnecessarily. If the PVO has at least one column configured as a Key column and one LastUpdateDate column configured as new data identifier, then set the load type to Incremental.
• Remove unwanted columns that are selected or enabled for replication.
• If the replication completes with warnings, check the error table from the target schema and make appropriate changes to the PVO configuration.
• Make sure that the PVO in the Oracle Fusion Cloud Applications data source has data records deleted frequently. If not, then clear the Include Deletions option.
• If a job fails or is cancelled, drop the staging and error table before running the job again.

Why does the time to run the same data replication differ on certain days?

The time that it takes to run a data replication job can vary due to various factors such as these:

• Oracle Autonomous Data Warehouse performance might affect the timings on a particular day.
• An Oracle Analytics Cloud instance running the replication job might be temporarily unavailable due to scheduled maintenance.

Is there a limit to the number of tables that I can add to a data replication job?

No, there's no limit to the number of tables that you can add to a job. You can execute no more than three replication jobs concurrently, but you can schedule any number of jobs concurrently. For example, three jobs can execute concurrently while other jobs are in the queue.

Is there a limit to the amount of data or number of rows that one data replication job can process?

No, a data replication job can process any amount of data or number of rows.

What other tips should I follow for data replication?

Follow these tips for data replication:

• Create fewer replications with more PVOs in each. Use the recommended extract PVOs.
• In the replication definition, deselect unwanted columns from the PVO.
• Use the "low" database service in Oracle Autonomous Data Warehouse for maximum concurrency.
• Schedule replication jobs to run at times where there's less load on Oracle Autonomous Data Warehouse.
• Keep the load type of the POVs set to the default, which is incremental mode.