diff options
Diffstat (limited to 'bin')
| -rwxr-xr-x | bin/pyspark | 1 | ||||
| -rwxr-xr-x | bin/spark-submit | 3 | ||||
| -rw-r--r-- | bin/spark-submit2.cmd | 3 |
3 files changed, 7 insertions, 0 deletions
diff --git a/bin/pyspark b/bin/pyspark index 776b28dc41..8acad61137 100755 --- a/bin/pyspark +++ b/bin/pyspark @@ -89,6 +89,7 @@ export PYTHONSTARTUP="$SPARK_HOME/python/pyspark/shell.py" if [[ -n "$SPARK_TESTING" ]]; then unset YARN_CONF_DIR unset HADOOP_CONF_DIR + export PYTHONHASHSEED=0 if [[ -n "$PYSPARK_DOC_TEST" ]]; then exec "$PYSPARK_DRIVER_PYTHON" -m doctest $1 else diff --git a/bin/spark-submit b/bin/spark-submit index bcff78edd5..0e0afe71a0 100755 --- a/bin/spark-submit +++ b/bin/spark-submit @@ -19,6 +19,9 @@ SPARK_HOME="$(cd "`dirname "$0"`"/..; pwd)" +# disable randomized hash for string in Python 3.3+ +export PYTHONHASHSEED=0 + # Only define a usage function if an upstream script hasn't done so. if ! type -t usage >/dev/null 2>&1; then usage() { diff --git a/bin/spark-submit2.cmd b/bin/spark-submit2.cmd index 08ddb18574..d3fc4a5cc3 100644 --- a/bin/spark-submit2.cmd +++ b/bin/spark-submit2.cmd @@ -20,6 +20,9 @@ rem rem This is the entry point for running Spark submit. To avoid polluting the rem environment, it just launches a new cmd to do the real work. +rem disable randomized hash for string in Python 3.3+ +set PYTHONHASHSEED=0 + set CLASS=org.apache.spark.deploy.SparkSubmit call %~dp0spark-class2.cmd %CLASS% %* set SPARK_ERROR_LEVEL=%ERRORLEVEL% |