ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jayush Luniya (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (AMBARI-9993) Add support for management of Phoenix Query Server to HDP Stack
Date Tue, 21 Apr 2015 18:28:58 GMT

    [ https://issues.apache.org/jira/browse/AMBARI-9993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14505441#comment-14505441
] 

Jayush Luniya commented on AMBARI-9993:
---------------------------------------

[~ndimiduk]
There were 2 approaches to going about adding phoenix query server as a component both would
work but each have their pros and cons that I will elaborate below

Approach 1: 
Create a separate common service for HBASE (v1.1.0.2.3) and make HDP 2.3 inherit from common-services/HBASE/1.1.0.2.3
So the inheritance would look like
{noformat}
common-services/HBASE/0.96.0.2.0
   stacks/HDP/2.0.6/services/HBASE
     stacks/HDP/2.1/services/HBASE
       stacks/HDP/2.2/services/HBASE
common-services/HBASE/1.1.0.2.3
   stacks/HDP/2.3/services/HBASE
{noformat}

Pros:
1. This would eliminate the need to add HDP2.3 checks in the code since HBASE/1.1.0.2.3 is
HDP2.3 and up anyways
2. Since the Phoenix Query Server is a new component in HBASE/1.1.0.2.3 logically it makes
sense that the script should reside here.
Cons:
1. This leads to a lot of code and config duplication if we fork the inheritance and would
incur more long term maintenance costs.

Approach 2: 
Add Phoenix Query Server to common-services/HBASE/0.96.0.2.0 and in stacks/HDP/2.3/services/HBASE/metainfo.xml
add this as a new component. 
Stack Inheritance
{noformat}
common-services/HBASE/0.96.0.2.0
   stacks/HDP/2.0.6/services/HBASE
     stacks/HDP/2.1/services/HBASE
       stacks/HDP/2.2/services/HBASE
         stacks/HDP/2.3/services/HBASE
{noformat}
Pros:
1. Since the package scripts folder is not duplicated here and the stack inheritance model
is not forked the maintenance cost is much lower.
Cons:
1. We will need to add stack version checks at certain places like for Ranger Solr support
in the scripts. 
2. Logically adding Phoenix Query Server script in HBASE/0.96.0.2.0 can lead to confusion.


What Andrew did was changed the solution to Approach 2. That leads to less code duplication
and shouldnt require the addendum patch. 

> Add support for management of Phoenix Query Server to HDP Stack
> ---------------------------------------------------------------
>
>                 Key: AMBARI-9993
>                 URL: https://issues.apache.org/jira/browse/AMBARI-9993
>             Project: Ambari
>          Issue Type: Improvement
>          Components: ambari-server
>    Affects Versions: 2.1.0
>            Reporter: Jeff Sposetti
>            Assignee: Nick Dimiduk
>             Fix For: 2.1.0
>
>         Attachments: 9993.jpg, 9993.patch, 9993.patch, AMBARI-9993.00.patch, AMBARI-9993.01.patch,
AMBARI-9993.02.patch, AMBARI-9993.03.patch, AMBARI-9993.04.patch, AMBARI-9993.05.patch, AMBARI-9993.addendum.patch,
AMBARI-9993.addendum.patch, AMBARI-9993.addendum.patch, AMBARI-9993.addendum.patch, displayname.jpg,
hostsummary.jpg, servicename.jpg, servicesummary.jpg
>
>
> Add support to install, manage and monitor the Phoenix Query Server.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message