[ 
https://issues.apache.org/jira/browse/CASSANDRA-13955?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16204625#comment-16204625
 ] 

Jason Brown commented on CASSANDRA-13955:
-----------------------------------------

I'm completely unfamiliar with the hadoop-related part of the code base, but as 
for resolving the NPE with {{DatabaseDescriptor#conf}}, can you call this 
somewhere in your code before initializing {{CqlBulkOutputFormat}}:

{code}
DatabaseDescriptor.toolInitialization()
{code}



> NullPointerException when using CqlBulkOutputFormat
> ---------------------------------------------------
>
>                 Key: CASSANDRA-13955
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-13955
>             Project: Cassandra
>          Issue Type: Bug
>         Environment: casandra-all:3.11.0
>            Reporter: Anish
>
> I am able to insert data of reducers by using CqlOutputFormat. Due to 
> performance reasons (I have large amount of inserts), I want to switch over 
> to CqlBulkOutputFormat. As per the documentation, everything remains same 
> except changing the format for reducers. But I get null pointer exception on 
> line below of CqlBulkRecordWriter
> {code}
> DatabaseDescriptor.setStreamThroughputOutboundMegabitsPerSec(Integer.parseInt(conf.get(STREAM_THROTTLE_MBITS,
>  "0")));
> {code}
> This is because "conf" is null in DatabaseDescriptor. I don't see any calls 
> where this would get initialized by reducer's CqlBulkOutputFormat
> Unfortunately, I could not find any documentation or samples of using 
> CqlBulkOutputFormat.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org

Reply via email to