Github user zentol commented on the pull request:
https://github.com/apache/flink/pull/1771#issuecomment-209400438
@rmetzger I hav efixed the issue you encountered. The problem was that the
JM tries to discard all state belonging to pending checkpoints upon job
failure, which effectively purged the write-ahead log. The simple fix is to not
do anything in ExactlyOnceState#discardState.
---
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.
---
