[
https://issues.apache.org/jira/browse/MRQL-73?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14557854#comment-14557854
]
Hudson commented on MRQL-73:
----------------------------
SUCCESS: Integrated in mrql-master-snapshot #20 (See
[https://builds.apache.org/job/mrql-master-snapshot/20/])
[MRQL-73] Set the max number of tasks in Spark mode (fegaras:
https://git-wip-us.apache.org/repos/asf?p=incubator-mrql.git&a=commit&h=5eb81d992cec29084d2a97686f95b08cdc727809)
* conf/mrql-env.sh
* bin/mrql.spark
> Set the max number of tasks in Spark mode
> -----------------------------------------
>
> Key: MRQL-73
> URL: https://issues.apache.org/jira/browse/MRQL-73
> Project: MRQL
> Issue Type: Bug
> Components: Run-Time/Spark
> Affects Versions: 0.9.6
> Reporter: Leonidas Fegaras
> Assignee: Leonidas Fegaras
> Priority: Critical
>
> The number of worker nodes in Spark distributed mode, which are specified by
> the MRQL -nodes parameter, must set the parameters SPARK_WORKER_INSTANCES
> (called SPARK_EXECUTOR_INSTANCES in Spark 1.3.*) and SPARK_WORKER_CORES;
> otherwise, Spark will always use all the available cores in the cluster.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
