[
https://issues.apache.org/jira/browse/SPARK-21370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16081436#comment-16081436
]
Apache Spark commented on SPARK-21370:
--------------------------------------
User 'brkyvz' has created a pull request for this issue:
https://github.com/apache/spark/pull/18595
> Clarify In-Memory State Store purpose (read-only, read-write) with an
> additional state
> --------------------------------------------------------------------------------------
>
> Key: SPARK-21370
> URL: https://issues.apache.org/jira/browse/SPARK-21370
> Project: Spark
> Issue Type: Improvement
> Components: Structured Streaming
> Affects Versions: 2.1.1
> Reporter: Burak Yavuz
> Assignee: Burak Yavuz
>
> Currently the HDFSBackedStateStore sets it's state as UPDATING as it is
> initialized.
> For every trigger, we create two state stores, one used during "Restore" and
> one during "Save". The "Restore" StateStore is read-only. This state store
> gets "aborted" after a task is completed, which results in a file being
> created and immediately deleted.
> This can be avoided if there is an INITIALIZED state and abort deletes files
> only when there is an update to the state store using "put" or "remove".
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
