Use the following rules and guidelines when you configure a Kafka session:
You cannot configure partitioning to read data from a Kafka source.
When you select JSON as the format type, you cannot read hierarchical data from a Kafka source or write hierarchical data to a Kafka target.
When you use a Kafka source or target definition in a mapping, you cannot configure multiple pipelines.
When you read data from a Kafka topic and write data to an Oracle target table, you must not select
Bulk
as the
Target Load Type
under the target session properties. Otherwise, the session fails.
When you read data from a Kafka topic or use a topic pattern and the offset of the last checkpoint is deleted during message recovery, you must provide the following property in the
Consumer Configuration Properties
to recover the messages from the next available offset:
auto.offset.reset=earliest
. Otherwise, the mapping reads data from the latest offset available.
When you remove the link for a port between the Kafka source definition and the Kafka Source Qualifier transformation, you must remove the port from the Kafka Source Qualifier transformation. Otherwise, the mapping fails with the following error:
com.informatica.powercenter.sdk.SDKException
When you read data from a Kafka topic and write data to any target, you must select
Resume from Last Checkpoint
as the
Recovery Strategy
.
When you write data to a Kafka target and enable message recovery for the session, the PowerCenter Integration Service only supports file-based recovery.
When you write data to a Kafka target and enable message recovery for the session, the PowerCenter Integration Service loads data from the last checkpoint stored in the recovery file.