[
https://issues.apache.org/jira/browse/HADOOP-17439?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17252894#comment-17252894
]
Ayush Saxena commented on HADOOP-17439:
---------------------------------------
Well I agree rolling back guava isn't an acceptable solution, that is why I
excluded that downgrade stuff earlier as well. But TBH this upgrade made
upgrading hadoop really tough....
These thirdparty libs if due to security reasons should be forced on all Apache
projects atleast. :(
Well the present problem of Hive(if it exists) isn't something we should try to
fix, By this logic if you keep hadoop libs always above hive libs in classpath,
then in that case leave guava, all common thirdparty libs shall make noise?
I tried running hive tests with the trunk version of hadoop, and most of them
which used to fail due to guava conflicts were passing.
[~leosun08] But if this classpath stuff exists, Isn't downstream very much
capable of solving this? For us getting completely rid of Guava isn't that
straightforward. Downgrading though on first shot looks good, but that is
unconventional. Can you point me the code where hive does this, putting hadoop
libs above its own in classpath or something related, from pom.xml's everything
looks at peace...
> No shade guava in trunk
> -----------------------
>
> Key: HADOOP-17439
> URL: https://issues.apache.org/jira/browse/HADOOP-17439
> Project: Hadoop Common
> Issue Type: Sub-task
> Reporter: Lisheng Sun
> Priority: Major
> Attachments: image-2020-12-18-22-01-45-424.png
>
>
> !image-2020-12-18-22-01-45-424.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
