Github user nathanmarz commented on the pull request:
https://github.com/apache/storm/pull/339#issuecomment-67581379
-1
Why are offset out of range exceptions happening when re-emitting a batch?
That shouldn't be possible so I'd like to know why this is happening.
Emitting a different batch than the first time actually breaks the contract
of transactional spouts â which is unacceptable (though it's fine for opaque
transactional spouts). Because it breaks the contract of transactional spouts,
it would be better to just error repeatedly than silently do the wrong thing.
The part of this code that catches the error and emits an empty batch for
the same batch id should only be applied towards the opaque spout.
Additionally, by better understanding why those offset out of range exceptions
are happening, there may be an alternative solution that doesn't involve
breaking the contract of transactional spouts.
---
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.
---
