Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
1.3.1, 1.4.0
-
None
Description
Currently, all Kafka consumer version still have this in the Javadoc:
The implementation currently accesses partition metadata when the consumer is constructed. That means that the client that submits the program needs to be able to reach the Kafka brokers or ZooKeeper.
This is also the case for the documentation:
The current FlinkKafkaConsumer implementation will establish a connection from the client (when calling the constructor) for querying the list of topics and partitions. For this to work, the consumer needs to be able to access the consumers from the machine submitting the job to the Flink cluster. If you experience any issues with the Kafka consumer on the client side, the client log might contain information about failed requests, etc.
These statements are no longer true since starting from Flink 1.3. partition metadata happens only in open().
Attachments
Issue Links
- links to