[
https://issues.apache.org/jira/browse/FLINK-16523?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
luojiangyu updated FLINK-16523:
-------------------------------
Description: The state for local recovery of the rocksdb backend. Using
the copy way for local state causes write IO pressure and write
amplification。Using HardLink way is more performance than Using copy way，
Whether replacing the copy way with hardlink is better
choice.(CheckpointStreamWithResultProvider.java) . [~richtesn] (was: The state
for local recovery of the rocksdb backend. Using the copy way for local state
causes write IO pressure and write amplification。Whether Using HardLink way is
more performance than Using copy way.(CheckpointStreamWithResultProvider.java)
. [~richtesn])
> The improvement for local state of rocksdb backend.
> ---------------------------------------------------
>
> Key: FLINK-16523
> URL: https://issues.apache.org/jira/browse/FLINK-16523
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / State Backends
> Affects Versions: 1.9.0, 1.9.1, 1.10.0
> Reporter: luojiangyu
> Priority: Major
>
> The state for local recovery of the rocksdb backend. Using the copy way for
> local state causes write IO pressure and write amplification。Using HardLink
> way is more performance than Using copy way， Whether replacing the copy way
> with hardlink is better choice.(CheckpointStreamWithResultProvider.java) .
> [~richtesn]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
