[
https://issues.apache.org/jira/browse/HIVE-13293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15280347#comment-15280347
]
Rui Li commented on HIVE-13293:
-------------------------------
Hi [~xuefuz], yeah order by is mostly at the end of stages. But that doesn't
mean the amount of data is small - that's why we need parallel order by. During
our benchmark, we hit OOM for several cases, which is due to some bug in Spark
1.6.0. So I thought using memory level cache may make it even worse.
To your second question, we unpersist cached RDDs at the end of each job. You
can refer to {{RemoteDriver#JobWrapper}} for that.
> Query occurs performance degradation after enabling parallel order by for
> Hive on Spark
> ---------------------------------------------------------------------------------------
>
> Key: HIVE-13293
> URL: https://issues.apache.org/jira/browse/HIVE-13293
> Project: Hive
> Issue Type: Bug
> Components: Spark
> Affects Versions: 2.0.0
> Reporter: Lifeng Wang
> Assignee: Rui Li
> Attachments: HIVE-13293.1.patch, HIVE-13293.1.patch
>
>
> I use TPCx-BB to do some performance test on Hive on Spark engine. And found
> query 10 has performance degradation when enabling parallel order by.
> It seems that sampling cost much time before running the real query.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
