Hi,

@chris @tdas


Referring to the latest integration documentation, it states the following:


"A single Kinesis input DStream can read from multiple shards of a Kinesis 
stream by creating multiple KinesisRecordProcessor threads."


But looking at the API and the example, each time we call 
KinesisUtils.createStream(), it is creating 1 dstream per shard and we should 
call it up to N shards per stream.


But according to documentation, we can increase the number of threads so 1 
dstream can read form multiple shards? Looking at the receiver code, there is 
only 1 thread and is not configurable?

https://github.com/apache/spark/blob/master/extras/kinesis-asl/src/main/scala/org/apache/spark/streaming/kinesis/KinesisReceiver.scala#L170


Thanks,

Liming



Reply via email to