Github user tillrohrmann commented on the issue:
https://github.com/apache/flink/pull/3278
I don't think that you have at least once guarantees if you disable
flushing. Assume the following: You have the input `event1, checkpoint barrier,
event2`. Now you write `event1` to Kafka but it is not yet committed. Now you
process the checkpoint barrier and complete the checkpoint without waiting for
`event1` to be written. Now writing `event1` to Kafka fails for some reason and
triggers a recovery. Then you will start at `event2` without having ever
written `event1` to Kafka.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
