Hi all. According to the document,
hdfs.rollCount is Number of events written to file before it rolled (0 = never roll based on number ) hdfs.batchSize is number of events written to file before it is flushed to HDFS Does hdfs sink append the batch data to the hdfs file if the file on hdfs not reached rollcount? If not what is the difference between hdfs.rollcount and hdfs.batchsize?
