[
https://issues.apache.org/jira/browse/HIVE-7076?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14006195#comment-14006195
]
Gunther Hagleitner commented on HIVE-7076:
------------------------------------------
Thanks [~vaibhavgumashta] and [~thejas]. On the 1.x front I propose the
following: Let's change the hive build to always build against hadoop 2.
Right now you can switch the dependency via -Phadoop-X flag, but that doesn't
really match what we ship. There's only one hive-exec jar etc not hive-exec-h1,
hive-exec-h2. What's more we're not catching binary incompatibilities the way
it's set up. Ideally we should build once (hadoop 2) then run the q file tests
with the -Phadoop-X flag. That way we will validate our build works correctly
against multiple hadoop versions.
[~thejas] thoughts?
Let me see if I can figure this out in mvn.
> Plugin (exec hook) to log to application timeline data to Yarn
> --------------------------------------------------------------
>
> Key: HIVE-7076
> URL: https://issues.apache.org/jira/browse/HIVE-7076
> Project: Hive
> Issue Type: Bug
> Reporter: Gunther Hagleitner
> Assignee: Gunther Hagleitner
> Attachments: HIVE-7076.1.patch
>
>
> See: https://issues.apache.org/jira/browse/YARN-1530
> This is a simple pre/post exec hook to log query + plan information to yarn.
> This information can be used to build tools and UIs to monitor, track, debug
> and tune Hive queries.
> Off by default, but can be enabled via:
> hive.exec.pre.hooks=ql.src.java.org.apache.hadoop.hive.ql.hooks.ATSHook
> hive.exec.post.hooks=ql.src.java.org.apache.hadoop.hive.ql.hooks.ATSHook
> hive.exec.failure.hooks=ql.src.java.org.apache.hadoop.hive.ql.hooks.ATSHook
--
This message was sent by Atlassian JIRA
(v6.2#6252)
