[
https://issues.apache.org/jira/browse/HBASE-21988?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16783722#comment-16783722
]
Sean Busbey commented on HBASE-21988:
-------------------------------------
bq. I thought it better to go the other way... keeping the bug fix and dropping
ancient versions of hadoop instead. Sean Busbey
I can understand the impulse, but we already established a way to do that:
dropping them in newer minor releases of HBase. if we can't even promise to
keep the same set of hadoop versions on a maintenance release? that's terrible
for the poor soul who can't change their Hadoop for whatever reason. How likely
that someone can update their Hadoop but not their HBase?
> Remove hadoop2.6.1-hadoop-2.6.4 as supported on branch-2.0
> ----------------------------------------------------------
>
> Key: HBASE-21988
> URL: https://issues.apache.org/jira/browse/HBASE-21988
> Project: HBase
> Issue Type: Sub-task
> Reporter: stack
> Assignee: stack
> Priority: Major
> Attachments: HBASE-21988.branch-2.0.001.patch
>
>
> See tail of parent issue where the parent patch breaks our compat for
> versions of hadoop before 2.6.4. Rather than remove the parent patch, will
> retract notion that branch-2.0 works on hadoop-2.6.1-4... Just leave 2.6.5 as
> supported.
> I'd remove support for hadoop-2.6 completely to align w/ what tip of branch-1
> says and what rest of branch-2 releases say but branch-2.0 has advertised
> itself up to this point as supporting hadoop-2.6 so requiring lastest release
> on that branch seems fine at this point, the EOL of branch-2.0
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
