Hi, we are using Kafka server + new client 0.9.0.1. In our application we process log events and aggregate the results in RDBMS. In order to achieve data consistency, when we process a polled log event list (a batch), we update DB in one transaction and we also save the partition offsets in the RDBMS within the same transaction. Then, on re-balance, we read the offsets from RDBMS.
So far so good. But we can see lot of auto-commit stuff in the log. So, we don't need the Kafka offset management. Is there a clean way to disable it? Is simple setting enable.auto.commit =false just enough? Isn't then still something remaining under the hood? I couldn't find some clean statement of how to setup kafka client when managing offsets on the application level in the doc or Internet. I found just this page: https://cwiki.apache.org/confluence/display/KAFKA/Offset+Management where the new offset management is discussed and the optionality is discussed just as an "interesting" thing, so that is basically what makes me unsure. Thank you Jakub Neubauer
