[
https://issues.apache.org/jira/browse/SPARK-33628?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17242951#comment-17242951
]
jinhai commented on SPARK-33628:
--------------------------------
And, Hive.getPartitions will cause MetaStoreClient close connection.
Similar issue: https://issues.apache.org/jira/browse/SPARK-29409
The chain of method calls is as follows:
Hive.getPartitions -> Hive.getUserName -> SessionState.getUserFromAuthenticator
-> SessionState.get().getAuthenticator -> SessionState.setupAuth ->
SessionState.setAuthorizerV2Config -> Hive.get -> Hive.getInternal ->
Hive.create -> Hive.close -> metaStoreClient.close
> Use the Hive.getPartitionsByNames method instead of Hive.getPartitions in the
> HiveClientImpl
> --------------------------------------------------------------------------------------------
>
> Key: SPARK-33628
> URL: https://issues.apache.org/jira/browse/SPARK-33628
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.0.0, 3.0.1
> Reporter: jinhai
> Priority: Major
> Attachments: image-2020-12-02-16-57-43-619.png
>
>
> When partitions are tracked by the catalog, that will compute all custom
> partition locations, especially when dynamic partitions, and the field
> staticPartitions is empty.
> The poor performance of the method listPartitions results in a long period
> of no response at the Driver.
> When read 12253 partitions, the method getPartitionsByNames takes 2 seconds,
> and the getPartitions takes 457 seconds, nearly 8 minutes
> !image-2020-12-02-16-57-43-619.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
