Can flink kafka producer dynamically discover new parititons after expansion??

2020-12-25 文章 ?????????? 
hi??I have a source Kafka and a sink Kafka??when the amount of data processing 
grows??I need to expand Kafka topic's  partition number ,but I don't want 
to restart the job to take effect.
for source Kafka, I use flink.partition-discovery.interval-millis and it could 
consume the new parititon after I expand the Kafka topic's partition 
number.


but sink kafka don't work like this??


The flink kafka producer get topic's paritition list and cache in 
topicPartitionsMap as showed In Class FlinkKafkaProducer

Can flink kafka producer dynamically discover new parititons after expansion??

2020-12-25 文章 ?????????? 
hi??I have a source Kafka and a sink Kafka??when the amount of data processing 
grows??I need to expand Kafka topic's  partition number ,but I don't want 
to restart the job to take effect.
for source Kafka, I use flink.partition-discovery.interval-millis and it could 
consume the new parititon after I expand the Kafka topic's partition 
number.


but sink kafka don't work like this??


The flink kafka producer get topic's paritition list and cache in 
topicPartitionsMap as showed In Class FlinkKafkaProducer