What happens when I partition data by key and then

问题:

What happens when I partition data by key and then later on add a new partition to the topic in Kafka?

Will there be any change to the existing record? And how will the future data be partitioned?

回答1:

Partitioning of existing data doesn't change when new partitions are added to a particular topic. Kafka will not attempt to re-distribute existing records and this modification will only have effect on new records. Note that by default, Kafka partitions data using hash(key) % noOfPartitions in order to ensure that records with the same key are added to the same partition. Data with different key will be added to partitions in a round-robin fashion.