[jira] [Commented] (HBASE-19982) Hadoop 2.8 and 2.9 Compatability

["Sean Busbey (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-19982?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16411687#comment-16411687 ] 

Sean Busbey commented on HBASE-19982:
-------------------------------------

quote from the "2.8 and 2.9 to precommit" jira:

{quote}Appy
Just adding to precommit doesn't give us anything - except that things compile, which can always be checked one-off.
Until we do the work of fully vetting a cluster against those versions, we can't call it supported. And if can't call them supported, there's no point of putting them in precommit.
If it was just 2-3 min overhead, i might have been onboard, but with ~10-15 min overhead, doesn't seem something that should be done.
{quote}

What would folks like to see as subtasks to accomplish this? (for ~1.5 timeframe maybe? Or 2.1?)

Bumping our default compilation / unit test version again? A job that ran UT/ITs against one of these versions? Someone signing up for manual testing (of what specifically)? Some number of production users?

FWIW I'm -1 on 2.9.0, since it's flagged as non-production. But if 2.9.1 was out and stable when the work started I'd want to aim for it, since it looks like the Hadoop community is phasing out maintenance releases on anything other than the latest minor release.

> Hadoop 2.8 and 2.9 Compatability
> --------------------------------
>
>                 Key: HBASE-19982
>                 URL: https://issues.apache.org/jira/browse/HBASE-19982
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Mike Drob
>            Priority: Major
>
> Hadoop 2.8.3 and 2.9.0 are both releases, and we haven't been keeping up with them.
> Several tasks we need to do here, will define them in sub tasks as we go along.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)