Very nice! I also was wondering about the offset autocommit in KafkaUtils. Since incoming streamed Kafka data is replicated across Spark nodes in memory it seems it is possible to have up to a batch of data loss if tasks hang or crash.
It seems you have avoided this case by using the Kafka simple consumer and managing the offsets explicitly. I think this will be a great addition to Spark Streaming, but curious what others think. -- View this message in context: http://apache-spark-user-list.1001560.n3.nabble.com/Low-Level-Kafka-Consumer-for-Spark-tp11258p11281.html Sent from the Apache Spark User List mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@spark.apache.org For additional commands, e-mail: user-h...@spark.apache.org
