Hi Upendra,
When both Change Tracking and Change Data Capture are enabled the connector generally prioritizes Change Tracking. Since CT requires the source table to have a primary key and table did not have one, it causes the pipeline to fail. Even after disabling CT, the same error may continue because the ingestion gateway caches the source database configuration and may take some time to detect changes. The gateway can take up to approximately 6 hours to refresh this information automatically. In addition, the pipeline may have existing streaming state or checkpoints that were established when CT was active.
On the SQL Server side, you can verify that CT has been disabled at both the database and table levels. You can also confirm that CDC is enabled at both the database and table levels. Once this is confirmed, restart the ingestion gateway. However, as the gateway may take time to pick up the updated configuration, the restart alone may not be sufficient to reinitialize the pipeline with CDC.
You need to perform a full refresh on the affected table. It forces the pipeline to re-snapshot the table and re-establish the change-tracking mechanism using the updated source configuration. You can trigger the full refresh from the pipeline UI by selecting Full Refresh for the specific table or use the Pipelines API with the parameters. This full refresh will take longer because the table needs to be re-snapshotted, but once it completes, incremental processing should resume using CDC.
You can try full refresh after switching from CT to CDC in dev. Confirm CT is disabled and CDC is enabled on SQL Server, restart the ingestion gateway, perform a full refresh for the affected table, monitor the pipeline update until it completes and then verify that incremental ingestion resumes using CDC. Switching to CDC can have a higher performance impact on the source database than CT, but it is necessary for a table without a primary key.
