3a281d1185
* shutting down the snapshotting thread and the DB history producer client if the connector is stopped while trying to write to the history topic * reducing the time that KafkaProducer#send() will block if Kafka isn't available; this will release the producer thread quicker in case the connector is stopped during snapshotting * not returning from finally block (!) in case the TX is rolled back; This prevented the failed state to be set by the outer catch clause in execute() |
||
---|---|---|
.. | ||
src | ||
pom.xml |