[
https://issues.apache.org/jira/browse/HIVE-8331?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14158740#comment-14158740
]
Xuefu Zhang commented on HIVE-8331:
-----------------------------------
+1
It seems that recent IOContext from trunk changed the the thread local property
of the class. The new way (initializing IOContext based on JobConf) doesn't
work for Spark. Thus, Spark will resort to the old way, maintain thread local
for IOContext. We can revisit this at a later stage.
> HIVE-8303 followup, investigate result diff [Spark Branch]
> ----------------------------------------------------------
>
> Key: HIVE-8331
> URL: https://issues.apache.org/jira/browse/HIVE-8331
> Project: Hive
> Issue Type: Task
> Components: Spark
> Reporter: Xuefu Zhang
> Assignee: Chao
> Attachments: HIVE-8331.1-spark.patch, HIVE-8331.2-spark.patch
>
>
> HIVE-8303 patch introduced some result diffs for some spark tests. We need to
> investigate those, including parallel_join0.q, union22.q,
> vectorized_shufflejoin.q, union_remove_18.q, and maybe more.
> Also the investigation includes the test failures related to spark.
> Specifically, union_remove_18.q demonstrated random order.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
