Zouxxyy commented on PR #8456:
URL: https://github.com/apache/hudi/pull/8456#issuecomment-1510835499
Currently the meaning of `clean.async.enabled` is ambiguous:
1. If it is used as whether enable asynchronous clean, then when it is
false, synchronous clean should not be executed in compact or cluster. Instead,
a parameter like `clean.async.enabled` should be added
```java
if (!conf.getBoolean(FlinkOptions.CLEAN_ASYNC_ENABLED)) {
LOG.info("Running inline clean");
this.writeClient.clean();
}
```
2. If it is used as the type of clean, then a parameter like
`clean.automatic` should be added to control whether to enable auto clean
Hudi spark is using the second logic, but for other table services of flink,
such as `clustering.async.enabled` and `compaction.async.enabled`, use the
first logic (as far as I understand flink does not support synchronous cluster
and compact, right?)
What do you think? @danny0405
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
