[
https://issues.apache.org/jira/browse/HBASE-11980?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141815#comment-14141815
]
Eric Charles commented on HBASE-11980:
--------------------------------------
[~stack] this is my +1 on
0001-HBASE-11980-Change-sync-to-hsync-remove-unused-InfoS.patch
I tested simple table manipulation in shell and ui, both on hadoop 2.5 and
hadoop 3.0.0-SNAPSHOT. I also reviewed the changes. All good.
A few comments:
1.- I need to run on jdk8 to avoid java.lang.NoSuchMethodError:
java.util.concurrent.ConcurrentHashMap.keySet()Ljava/util/concurrent/ConcurrentHashMap$KeySetView.
Normal ?
2.- Buidling with default profile (2) works well for hadoop-3 - this is because
2.4 is close from 3-snapshot). What a voodoo, building with 2.4, runs on 2.5
and 3-snapshot... Still this is confusing for end-user.
3.- I didn't test with profile hadoop-3, only with default one.
For the simplicity, I would simply remove hadoop-3, and rename 'hadoop-2' to
'hadoop-current' or something like that, or do you expect future divergence
between hadoop2 and hadoop3?
> Change sync to hsync, remove unused InfoServer, and reference our httpserver
> instead of hadoops
> -----------------------------------------------------------------------------------------------
>
> Key: HBASE-11980
> URL: https://issues.apache.org/jira/browse/HBASE-11980
> Project: HBase
> Issue Type: Sub-task
> Reporter: stack
> Assignee: stack
> Fix For: 2.0.0, 0.99.1
>
> Attachments:
> 0001-HBASE-11980-Change-sync-to-hsync-remove-unused-InfoS.patch
>
>
> Subtask that fixes a few issues with our building against hadoop3. Get it in.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
