question about TransactionalTridentKafkaSpout

[Weide Zhang <we...@gmail.com>]

Hi,

I'm new to storm and trident. I have some general questions regarding usage
of TransactionTridentKafkaSpout. I'm trying to understand the intervals but
not sure where can i find some good resource.

1. Say one kafka topic has n partitions, how does storm worker/executor
mapped to partitions ? Is it one to one mapping or one to many mapping  ?

2. how does the tuple batch work for kafka that has n partitions ? Does the
batching across all partitions or each partitions has its own batch ?

3. say one of the worker that emits one of the topic partitions data dies,
how does the recovery mechanism work ?

4. if new broker / partition is dynamically get changed, can storm
dynamically adjust to that ?

Thanks a lot,

Weide