[ 
https://issues.apache.org/jira/browse/SPARK-17814?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

shreyas subramanya updated SPARK-17814:
---------------------------------------
    Description: 
One of our customers is trying to pass in json through spark-submit as follows:
spark-submit --verbose --class SimpleClass --master yarn-cluster ./simple.jar 
"{\"mode\":\"wf\", \"arrays\":{\"array\":[1]}}"

When this argument reaches the application, it is truncated in yarn-cluster 
mode as follows:
Passed args in yarn-cluster mode: {"mode":"wf", "arrays":{"array":[1]

The spark environment from history server also confirms this:
org.apache.spark.deploy.yarn.ApplicationMaster --class SimpleClass --jar 
file:/home/user/simple.jar --arg {"mode":"wf", "arrays":{"array":[1] 
--executor-memory 1024m --executor-cores 1 --properties-file 
/data/usercache/user/appcache/application_1475790979221_0010/container_1475790979221_0010_02_000001/__spark_conf__/__spark_conf__.properties

If the same application is submitted in yarn-client mode, it works. The 
application prints:
Passed args in yarn-client mode: {"mode":"wf", "arrays":{"array":[1]}}

  was:
One of our customers is trying to pass in json through spark-submit as follows:
spark-submit --verbose --class SimpleClass --master yarn-cluster ./simple.jar 
"{\"mode\":\"wf\", \"arrays\":{\"array\":[1]}}"

When this argument reaches the application, it is truncated in yarn-cluster 
mode as follows:
Passed args in yarn-client mode: {"mode":"wf", "arrays":{"array":[1]}}
Passed args in yarn-cluster mode: {"mode":"wf", "arrays":{"array":[1]

The spark environment from history server also confirms this:
org.apache.spark.deploy.yarn.ApplicationMaster --class SimpleClass --jar 
file:/home/user/simple.jar --arg {"mode":"wf", "arrays":{"array":[1] 
--executor-memory 1024m --executor-cores 1 --properties-file 
/data/usercache/user/appcache/application_1475790979221_0010/container_1475790979221_0010_02_000001/__spark_conf__/__spark_conf__.properties


> spark submit arguments are truncated in yarn-cluster mode
> ---------------------------------------------------------
>
>                 Key: SPARK-17814
>                 URL: https://issues.apache.org/jira/browse/SPARK-17814
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Submit, YARN
>    Affects Versions: 1.6.1
>            Reporter: shreyas subramanya
>
> One of our customers is trying to pass in json through spark-submit as 
> follows:
> spark-submit --verbose --class SimpleClass --master yarn-cluster ./simple.jar 
> "{\"mode\":\"wf\", \"arrays\":{\"array\":[1]}}"
> When this argument reaches the application, it is truncated in yarn-cluster 
> mode as follows:
> Passed args in yarn-cluster mode: {"mode":"wf", "arrays":{"array":[1]
> The spark environment from history server also confirms this:
> org.apache.spark.deploy.yarn.ApplicationMaster --class SimpleClass --jar 
> file:/home/user/simple.jar --arg {"mode":"wf", "arrays":{"array":[1] 
> --executor-memory 1024m --executor-cores 1 --properties-file 
> /data/usercache/user/appcache/application_1475790979221_0010/container_1475790979221_0010_02_000001/__spark_conf__/__spark_conf__.properties
> If the same application is submitted in yarn-client mode, it works. The 
> application prints:
> Passed args in yarn-client mode: {"mode":"wf", "arrays":{"array":[1]}}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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

Reply via email to