[
https://issues.apache.org/jira/browse/AMBARI-7229?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14127697#comment-14127697
]
Jaimin D Jetly commented on AMBARI-7229:
----------------------------------------
[~adenisso]
The current trunk for ambari-1.7.0 release has already accomplished what you
have proposed (defining all properties in stack instead of defining it in
javascript depending upon the stack version)
Can you please point out few properties as example that are distinguished and
processed by javascript code only for Hadoop 2.x stack so that we can verify
that the trunk code already does what you are proposing ?
> Stack version must start with 2. to be considered Hadoop-2.x compatible
> -----------------------------------------------------------------------
>
> Key: AMBARI-7229
> URL: https://issues.apache.org/jira/browse/AMBARI-7229
> Project: Ambari
> Issue Type: Bug
> Affects Versions: 1.6.1
> Reporter: Alexander Denissov
>
> We have defined BIGTOP 0.8 stack based on upcoming BigTop 0.8 release. Stack
> definition is similar to HDP-2.1 merged with HDP-2.0.6. When this stack is
> configured for deployment a few of HDFS property defaults are missing as
> indicated by red boxes.
> Looking at JavaScript logic is seems these properties are processed only for
> Hadoop 2.x compatible stacks and App.js checks if the stack version starts
> with 2 to consider it Hadoop 2.x compatible, which is not the case with 0.8
> version of BIGTOP stack.
> PROPOSAL: the fact that a given stack is Hadoop 2.x compatible should not be
> tied to the vendor stack version number, but rather explicitly defined as a
> property either in the stack definition or the service definition.
> Alternatively, looking at the actual version of the HDFS service included in
> the stack will be able to reveal if the stack is Hadoop 2.x compliant.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
