[ 
https://issues.apache.org/jira/browse/IMPALA-15103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18089384#comment-18089384
 ] 

ASF subversion and git services commented on IMPALA-15103:
----------------------------------------------------------

Commit 33d9560f52717994dc7948756c571f65e6338df2 in impala's branch 
refs/heads/master from Joe McDonnell
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=33d9560f5 ]

IMPALA-15103: Hive configuration should honor the HDFS_REPLICATION env variable

This changes hive-site.xml.py to use the HDFS_REPLICATION
environment variable to set Hive's configuration for
HDFS replication. This makes setting HDFS_REPLICATION=1
apply to Hive tables. This is particularly useful when
using larger scale factors of TPC-H/TPC-DS to avoid excess
disk space use. Using this on a standard dataload is
fairly functional, but it will break some planner tests.

Testing:
 - Ran bin/single_node_perf_run.py with this setting and
   monitored disk space usage

Change-Id: Iceeaed41eb5b3797459c0a3035d159d903453eeb
Reviewed-on: http://gerrit.cloudera.org:8080/24461
Reviewed-by: Joe McDonnell <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>


> Minicluster Hive should respect the HDFS_REPLICATION env variable
> -----------------------------------------------------------------
>
>                 Key: IMPALA-15103
>                 URL: https://issues.apache.org/jira/browse/IMPALA-15103
>             Project: IMPALA
>          Issue Type: Task
>          Components: Infrastructure
>    Affects Versions: Impala 5.0.0
>            Reporter: Joe McDonnell
>            Assignee: Joe McDonnell
>            Priority: Major
>
> HDFS replication defaults to 3, which dramatically increases disk space for 
> tables. We have an existing HDFS_REPLICATION environment variable, but Hive 
> does not use it in its configurations. We should change Hive's configurations 
> to respect this environment variable. This is particularly useful for 
> perf-AB-test jobs where we are loading large scale TPC-H or TPC-DS and want 
> to avoid the disk space use.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to