[
https://issues.apache.org/jira/browse/HIVE-21524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16803907#comment-16803907
]
David Mollitor commented on HIVE-21524:
---------------------------------------
It would allow some level of unification. The authorization and auditing could
all occur in one place, it allows the users to only have to worry about one
connection string, and as I understand it, Impala does not consider Hive locks
when it is performing its actions. Tunneling the queries through HS2 first
could apply the appropriate locks before passing query to Impala.
> Impala Engine
> -------------
>
> Key: HIVE-21524
> URL: https://issues.apache.org/jira/browse/HIVE-21524
> Project: Hive
> Issue Type: New Feature
> Affects Versions: 4.0.0
> Reporter: David Mollitor
> Priority: Major
>
> Now that Impala has "dedicated coordinator" capability, it could be
> interesting to pair HiveServer2 instances with Impala dedicated coordinators
> on the same localhost. A client could request an 'impala' execution engine
> and subsequent queries would be routed to the local coordinator.
> {code:sql}
> set hive.execution.engine=impala;
> {code}
> This would allow clients seamless access to both capabilities without needing
> different connections or drivers, Hive would also be a central location for
> auditing and authorization.
> https://www.cloudera.com/documentation/enterprise/latest/topics/impala_dedicated_coordinator.html
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
