You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Andrew Or (JIRA)" <ji...@apache.org> on 2014/05/21 06:11:38 UTC
[jira] [Updated] (SPARK-1894) The default ec2 set-up ignores
--driver-class-path
[ https://issues.apache.org/jira/browse/SPARK-1894?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Andrew Or updated SPARK-1894:
-----------------------------
Description:
If the user sets up an ec2 cluster, the scripts automatically add the following lines to conf/spark-env.sh
{code}
export SPARK_SUBMIT_LIBRARY_PATH="/root/ephemeral-hdfs/lib/native/"
export SPARK_SUBMIT_CLASSPATH=$SPARK_CLASSPATH:/root/ephemeral-hdfs/conf"
{code}
Unfortunately, these variables are exported after spark-submit parses the --driver-* flags, which also set these variables. As a result, all values set via the --driver-* flags get overridden.
The simple fix is to append to, instead of overwrite, these variables.
was:
If the user sets up an ec2 cluster, the scripts automatically add the following lines to conf/spark-env.sh
<code>
export SPARK_SUBMIT_LIBRARY_PATH="/root/ephemeral-hdfs/lib/native/"
export SPARK_SUBMIT_CLASSPATH=$SPARK_CLASSPATH:/root/ephemeral-hdfs/conf"
<code>
Unfortunately, these variables are exported after spark-submit parses the --driver-* flags, which also set these variables. As a result, all values set via the --driver-* flags get overridden.
The simple fix is to append to, instead of overwrite, these variables.
> The default ec2 set-up ignores --driver-class-path
> --------------------------------------------------
>
> Key: SPARK-1894
> URL: https://issues.apache.org/jira/browse/SPARK-1894
> Project: Spark
> Issue Type: Bug
> Affects Versions: 1.0.0
> Reporter: Andrew Or
> Fix For: 1.0.1
>
>
> If the user sets up an ec2 cluster, the scripts automatically add the following lines to conf/spark-env.sh
> {code}
> export SPARK_SUBMIT_LIBRARY_PATH="/root/ephemeral-hdfs/lib/native/"
> export SPARK_SUBMIT_CLASSPATH=$SPARK_CLASSPATH:/root/ephemeral-hdfs/conf"
> {code}
> Unfortunately, these variables are exported after spark-submit parses the --driver-* flags, which also set these variables. As a result, all values set via the --driver-* flags get overridden.
> The simple fix is to append to, instead of overwrite, these variables.
--
This message was sent by Atlassian JIRA
(v6.2#6252)