After spark converged with Hadoop neither Spark.yarn.jars nor Spark.yarn.archive is set

Reference documents:

http://blog.csdn.net/lxhandlbb/article/details/54410644

Each time a spark task is submitted to yarn, uploading resource (package spark jars and upload) is always present on HDFs.

In bad condition, it will be stuck here for a long time.

Solve:
To create a directory on HDFs:
HDFs Dfs-mkdir/spark_jars

Upload Spark's jars (spark1.6 only need to upload Spark-assembly-1.6.0-snapshot-hadoop2.6.0.jar)

HDFs dfs-put/opt/spark/jars/*/spark_jars

In Spark's Conf's spark-default.conf

To add a configuration:

spark.yarn.jars=hdfs://master:9000/spark_jars/*
#spark. yarn.jars=/opt/spark-2.1.1-bin-hadoop2.7/jars/*

can be resolved. This problem does not occur.

After spark converged with Hadoop neither Spark.yarn.jars nor Spark.yarn.archive is set