amazon web services - How to prevent EMR Spark step from retrying? -
i have aws emr cluster (emr-4.2.0, spark 1.5.2), submitting steps aws cli. problem is, if spark application fails, yarn trying run application again (under same emr step). how can prevent this?
i trying set --conf spark.yarn.maxappattempts=1
, correctly set in environment/spark properties, doesn't prevent yarn restarting application.
you should try set spark.task.maxfailures
1 (4 default).
meaning:
number of failures of particular task before giving on job. total number of failures spread across different tasks not cause job fail; particular task has fail number of attempts. should greater or equal 1. number of allowed retries = value - 1.
Comments
Post a Comment