Github user shikhar commented on the pull request:
https://github.com/apache/flink/pull/1690#issuecomment-207563659
Another thing is that presence of offsets in ZK (0.8) / Kafka (0.9) also
affects what happens when a job is starting without a checkpoint / savepoint.
If they are present, the reset mode won't be respected, and there is potential
for dropping messages (e.g. if the recovery strategy in the absence of a valid
checkpoint or savepoint is to process from the beginning of available data with
offset reset mode 'earliest').
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
