[
https://issues.apache.org/jira/browse/SPARK-45504?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-45504:
-----------------------------------
Labels: pull-request-available (was: )
> RocksDB State Store Should Lower RocksDB Background Thread CPU Priority
> -----------------------------------------------------------------------
>
> Key: SPARK-45504
> URL: https://issues.apache.org/jira/browse/SPARK-45504
> Project: Spark
> Issue Type: Task
> Components: Structured Streaming
> Affects Versions: 3.5.1
> Reporter: Siying Dong
> Priority: Minor
> Labels: pull-request-available
>
> We can move RocksDB flush and compaction to lower CPU priority. They usually
> are background tasks and don't need to compete with task execution. For the
> case where a task may wait for some RocksDB background task to finish, such
> as checkpointing, or waiting async checkpointing to finish, the task slot is
> waiting so we are likely to have enough CPU for low pri CPU anyway.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
