You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Natalino Busa (JIRA)" <ji...@apache.org> on 2019/04/10 10:37:00 UTC
[jira] [Created] (SPARK-27427) cannot change list of packages from
PYSPARK_SUBMIT_ARGS after restarting context
Natalino Busa created SPARK-27427:
-------------------------------------
Summary: cannot change list of packages from PYSPARK_SUBMIT_ARGS after restarting context
Key: SPARK-27427
URL: https://issues.apache.org/jira/browse/SPARK-27427
Project: Spark
Issue Type: Bug
Components: PySpark
Affects Versions: 2.4.0
Reporter: Natalino Busa
Upon creating the gateway the env variable PYSPARK_SUBMIT_ARGS , if packages are provided the jars will be downloaded and added to the list of files to distribute to the cluster.
However this mechanism only works once because the gateway is kept, when the SparkContext is stopped and created again. Possible root cause the gateway is only created once.
This more aggressive SparkSession stop forces the gateway to be recreated
```
def stop(spark_session=None):
try:
sc = None
if spark_session:
sc = spark_session.sparkContext
spark_session.stop()
cls = pyspark.SparkContext
sc = sc or cls._active_spark_context
if sc:
sc.stop()
sc._gateway.shutdown()
cls._active_spark_context = None
cls._gateway = None
cls._jvm = None
except Exception as e:
print(e)
logging.warning('Could not fully stop the engine context')
```
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org