[jira] [Commented] (HIVE-15678) Hive On Spark queries fail when HBase is configured and down even when the query doesn't rely on HBase

["Xuefu Zhang (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-15678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15831733#comment-15831733 ] 

Xuefu Zhang commented on HIVE-15678:
------------------------------------

Hive on Spark doesn't attempt to make any HBase connection. I vaguely remember the behavior was observed in Spark in the past. Nevertheless, if you don't plan to access HBase, it might make sense to unset these properties.

> Hive On Spark queries fail when HBase is configured and down even when the query doesn't rely on HBase
> ------------------------------------------------------------------------------------------------------
>
>                 Key: HIVE-15678
>                 URL: https://issues.apache.org/jira/browse/HIVE-15678
>             Project: Hive
>          Issue Type: Improvement
>          Components: HiveServer2
>            Reporter: Thomas Scott
>            Priority: Minor
>
> In a scenario where the Hive execution engine is spark and the following properties are configured in hive-site.xml:
> hbase.zookeeper.quorum
> hbase.zookeeper.property.clientPort
> Hive On Spark queries fail if HBase is down as they are trying to connect to the HBase master. This occurs even when the queries issued to not need HBase to complete.
> Can the HBase connection only be made on demand if the query requires it?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)