Want to ask three questions about the connector to kafka!

[大吃一惊 <md...@qq.com>]

background: I want to use the flink to get the log stream once and according to different demand to add or remove some params in each line from kafka, then distribute to different kafka cluster or different topic.

questions: The attachment is the simply example I used to test distribute to different topic, I want to know
1) I want to read stream only once, not get many times from src kafka, dose it code work? or How can I do that?
2) Does each result stream independent, not block each others?
3) It is the end-to-end mode, how to solve the excatly-once problem, can only use checkpoint work?


first time to use kafka, this is large scale logging analysis system, thank you for you help!