Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Wed, 8 Nov 2017 06:25:01 +0000 (UTC)
From: "Yung-An He (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.13116680.1510062865000.178620.1510122301069@Atlassian.JIRA>
In-Reply-To: <JIRA.13116680.1510062865000@Atlassian.JIRA>
References: <JIRA.13116680.1510062865000@Atlassian.JIRA> <JIRA.13116680.1510062865080@jira-lw-us.apache.org>
Subject: [jira] [Commented] (HBASE-19201) BulkLoading in HBaseContext in
 hbase-spark does not close connection
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Wed, 08 Nov 2017 06:44:04 -0000


    [ https://issues.apache.org/jira/browse/HBASE-19201?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16243432#comment-16243432 ] 

Yung-An He commented on HBASE-19201:
------------------------------------

Hi

Could you post the needed code which you copied into the scala project? 

> BulkLoading in HBaseContext in hbase-spark does not close connection
> --------------------------------------------------------------------
>
>                 Key: HBASE-19201
>                 URL: https://issues.apache.org/jira/browse/HBASE-19201
>             Project: HBase
>          Issue Type: Bug
>          Components: hbase
>    Affects Versions: 1.1.12
>         Environment: I was using the cdh 5.11.1 version but I checken on newest branch and problem persists
>            Reporter: Lucas Resch
>              Labels: newbie
>   Original Estimate: 2h
>  Remaining Estimate: 2h
>
> Within the hbase-spark module an HBaseContext exists that provides utility functions to do bulkLoading data in HBase. I tried using this function in a streaming context, but after a while Zookeeper denies further connections since the maximum of connections per client is exhausted. 
> This issue seems to be within HBaseContext, since the functions bulkLoad and bulkLoadThinRows open a connection via the ConnectionFactory, but never closes that connection.
> I copied the needed code into a new scala project and added a conn.close() at the end of the function and the problem is gone. 
> It seems like no one else has had this problem before. I'm guessing thats because almost no one uses its function within a streaming context. And a one time call to it with RDDs might never reach that upper limit on connections. 


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)