Github user aljoscha commented on the issue:
https://github.com/apache/flink/pull/4301
Note, that this doesn't normally occur because the strategy for assigning
Kafka partitions and for assigning operator state is the same (right now).
However, this means that you will have active partition discovery for parallel
instances that didn't previously have state: assume we have 1 partition and 1
parallel source. Now we add a new partition and restart the Flink job. Now,
parallel instance 1 will still read from partition 0, parallel instance 2 will
think that it didn't restart (because it didn't get state) and will discover
partitions and take ownership of partition 1.
(This is with current `release-1.3` branch code.)
---
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.
---
