I am trying to run the JavaKinesisWordCountASL example.
The example seem to connect to my Kinesis Stream and gets data from the stream (as shown in the log below).
This might have something to do with how many worker thread you got. I had the same problem when I ran the app with --master local[2]. I spent numerous hours searching for an answer and found nothing. Just out of curiosity, I changed to --master local[4] and it worked. I do not know the root cause. Maybe somebody more familiar with Spark can enlighten us.
Note: in my case, my Kinesis stream had two shards. So the app created two input streams, one for each shard.