[ 
https://issues.apache.org/jira/browse/KYLIN-3168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16337797#comment-16337797
 ] 

wuyingjun edited comment on KYLIN-3168 at 1/25/18 9:31 AM:
-----------------------------------------------------------

[~Shaofengshi]

In the hadoop environment, the hbase cluster  has accss control; The app user 
can not get the right hbase configuration like quorum just using the following 
code.
 HBaseConfiguration.create(getConf())
  

Actually, Simple creating hbase configuration is not right. Since we can get 
the global hbase configuration why not using it ? 

The initialization of the current Hbase configuration has runned the following 
code, we can figure it out in the attachement.
 HBaseConnection.addHBaseClusterNNHAConfiguration(job.getConfiguration());
  

!飞信截图20180124232443.png!


was (Author: wuyingjun):
In the hadoop environment, the hbase cluster  has accss control; The app user 
can not get the right hbase configuration like quorum just using the following 
code.
HBaseConfiguration.create(getConf())
 

Actually, Simple creating hbase configuration is not right. Since we can get 
the global hbase configuration why not using it ? 

The initialization of the current Hbase configuration has runned the following 
code, we can figure it out in the attachement.
HBaseConnection.addHBaseClusterNNHAConfiguration(job.getConfiguration());
 

!飞信截图20180124232443.png!

> CubeHFileJob should use currentHBaseConfiguration but not new create hbase 
> configuration
> ----------------------------------------------------------------------------------------
>
>                 Key: KYLIN-3168
>                 URL: https://issues.apache.org/jira/browse/KYLIN-3168
>             Project: Kylin
>          Issue Type: Bug
>    Affects Versions: v2.2.0
>            Reporter: wuyingjun
>            Assignee: wuyingjun
>            Priority: Major
>         Attachments: AfterModified.png, CubeHFileJob_Exception.png, 
> KYLIN-3168.patch, 飞信截图20180124232443.png
>
>
> CubeHFileJob  use new hbase configuration is not correct because when the 
> zookeeper quorum is not localhost (job worker), kylin may account zookeeper 
> ConnectionLossException.
>  



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

Reply via email to