Harish created SPARK-20362:
------------------------------
Summary: spark submit not considering user defined Configs
(Pyspark)
Key: SPARK-20362
URL: https://issues.apache.org/jira/browse/SPARK-20362
Project: Spark
Issue Type: Bug
Components: PySpark
Affects Versions: 2.1.0
Reporter: Harish
I am trying to set up the custom configuration on runtime (pyspark), but in my
spark UI <ip>:8080 i see my job is using complete node/cluster resources and
application name is "test.py"(which is script name). It looks like the user
defined configurations are not considered in job submit.
command : spark-submit test.py
standalone mode(2 nodes and 1 master)
Here is the code:
test.py
from pyspark.sql import SparkSession
from pyspark import SparkConf
if __name__ == "__main__":
conf = SparkConf().setAll([('spark.executor.memory', '8g'),
('spark.executor.cores', '3'), ('spark.cores.max', '10'),
('spark.driver.memory','8g')])
spark =
SparkSession.builder.config(conf=conf).enableHiveSupport().getOrCreate()
sc = spark.sparkContext
print(sc.getConf().getAll())
sqlContext = SQLContext(sc)
hiveContext = HiveContext(sc)
print(hiveContext)
print(sc.getConf().getAll())
print("Complete")
Print:
[('spark.jars.packages', 'com.databricks:spark-csv_2.11:1.2.0'),
('spark.local.dir', '/mnt/sparklocaldir/'), ('hive.metastore.warehouse.dir',
'<path>'), ('spark.app.id', 'app-20170417221942-0003'), ('spark.jars',
'file:/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,file:/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,file:/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.executor.id', 'driver'), ('spark.app.name', 'test.py'),
('spark.cores.max', '10'), ('spark.serializer',
'org.apache.spark.serializer.KryoSerializer'), ('spark.driver.port', '35596'),
('spark.sql.catalogImplementation', 'hive'), ('spark.sql.warehouse.dir',
'<path>'), ('spark.rdd.compress', 'True'), ('spark.driver.memory', '8g'),
('spark.serializer.objectStreamReset', '100'), ('spark.executor.memory', '8g'),
('spark.executor.cores', '3'), ('spark.submit.deployMode', 'client'),
('spark.files',
'file:/home/user/test.py,file:/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,file:/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,file:/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.master', 'spark://master:7077'), ('spark.submit.pyFiles',
'/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.driver.host', 'master')]
<pyspark.sql.context.HiveContext object at 0x7f6f87b2e5f8>
[('spark.jars.packages', 'com.databricks:spark-csv_2.11:1.2.0'),
('spark.local.dir', '/mnt/sparklocaldir/'), ('hive.metastore.warehouse.dir',
'<path>'), ('spark.app.id', 'app-20170417221942-0003'), ('spark.jars',
'file:/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,file:/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,file:/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.executor.id', 'driver'), ('spark.app.name', 'test.py'),
('spark.cores.max', '10'), ('spark.serializer',
'org.apache.spark.serializer.KryoSerializer'), ('spark.driver.port', '35596'),
('spark.sql.catalogImplementation', 'hive'), ('spark.sql.warehouse.dir',
'<path>'), ('spark.rdd.compress', 'True'), ('spark.driver.memory', '8g'),
('spark.serializer.objectStreamReset', '100'), ('spark.executor.memory', '8g'),
('spark.executor.cores', '3'), ('spark.submit.deployMode', 'client'),
('spark.files',
'file:/home/user/test.py,file:/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,file:/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,file:/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.master', 'spark://master:7077'), ('spark.submit.pyFiles',
'/home/user/.ivy2/jars/com.databricks_spark-csv_2.11-1.2.0.jar,/home/user/.ivy2/jars/org.apache.commons_commons-csv-1.1.jar,/home/user/.ivy2/jars/com.univocity_univocity-parsers-1.5.1.jar'),
('spark.driver.host', 'master')]
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
