79546211

Date: 2025-03-31 12:26:50
Score: 0.5
Natty:
Report link

After modifying the following I was able to run in cluster deploy mode on my standalone cluster. The job that finishes without errors in Client Deploy Mode is now failing with a typical distribution issue (serialization of a UDF)

Environment settings -> deploy mode cluster

Stdout

removed SPARK_LOCAL_IP from the spark-env.sh
removed spark.driver.host from spark-default.conf
spark.driver.bindAddress=<hostname of the node on which spark-default.conf file is modified on>
  I have three nodes and set each of their files to their own host name

(note : removed spark.executor.host as it doesn't even exist :-S )

Reasons:
  • Long answer (-0.5):
  • Has code block (-0.5):
  • Self-answer (0.5):
  • Low reputation (1):
Posted by: Rene