Closed willb closed 8 years ago
MLLib in PySpark depends on NumPy, so we'll need to make sure the workers have access to it -- either via the openshift-spark image or via the pyspark image here (distributing NumPy to workers with the --py-files option to spark-submit).
openshift-spark
--py-files
spark-submit
Fixed in
https://github.com/radanalyticsio/oshinko-s2i/pull/40
MLLib in PySpark depends on NumPy, so we'll need to make sure the workers have access to it -- either via the
openshift-spark
image or via the pyspark image here (distributing NumPy to workers with the--py-files
option tospark-submit
).