[
https://issues.apache.org/jira/browse/FLINK-28812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17638073#comment-17638073
]
Jingsong Lee commented on FLINK-28812:
--------------------------------------
master:
638c34b39ab0b57b14cef9b050c749de7b1bf7d3
> Add read.compacted to read compacted snapshot only
> --------------------------------------------------
>
> Key: FLINK-28812
> URL: https://issues.apache.org/jira/browse/FLINK-28812
> Project: Flink
> Issue Type: Sub-task
> Reporter: Jingsong Lee
> Priority: Major
> Labels: pull-request-available
>
> If we separate the compaction from the writing, perhaps a large number of
> level0 files will be generated, which may lead to an OOM of merging too many
> files when reading.
> We can introduce `read.compacted` to avoid reading uncompacted snapshot.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
