Github user rmetzger commented on the pull request:
https://github.com/apache/flink/pull/1690#issuecomment-199316872
Let me summarize the change:
```
auto.commit.enable=true checkpointing=true offsets go to ZK/Broker on
each completed checkpoint
auto.commit.enable=false checkpointing=true offsets are only available
for recovery with Flink, not in ZK / broker
auto.commit.enable=true checkpointing=false offsets are written to ZK
periodically
auto.commit.enable=false checkpointing=false offsets are not in ZK or in
Flink
```
I don't like this variant:
```
auto.commit.enable=false checkpointing=true offsets are only available
for recovery with Flink, not in ZK / broker
```
because it stretches the original definition of "auto.commit.enable" a bit.
In my understanding, "auto.commit.enable" in Kafka starts a periodic committer,
for sending the offsets to ZK / broker.
In Flink, the semantic of the setting is more or less "enable offset
committing".
How about we limit the "auto.commit.enable" setting only to the periodic
offset committing thread?
---
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.
---
