[转] Spark-Sql On YARN自动调整Executor数配置

在所有的NodeManager中，修改yarn-site.xml，为yarn.nodemanager.aux-services添加spark_shuffle值，并设置yarn.nodemanager.aux-services.spark_shuffle.class值为org.apache.spark.network.yarn.YarnShuffleService，如下：

<property>

<name>yarn.nodemanager.aux-services</name>

<value>mapreduce_shuffle,spark_shuffle<value>

</property>

<property>

<name>yarn.nodemanager.aux-services.spark_shuffle.class</name>

<value>org.apache.spark.network.yarn.YarnShuffleService</value>

</property>

配置 $SPARK_HOME/conf/spark-default.xml，添加以下两项

spark.dynamicAllocation.minExecutors 1 #最小Executor数

spark.dynamicAllocation.maxExecutors 100 #最大Executor数 

执行时开启自动调整Executor数开关，以spark-sql yarn client模式为例

spark-submit \

--class SySpark.SqlOnSpark \

--master yarn-client \

--conf spark.shuffle.service.enabled=true \

--conf spark.dynamicAllocation.enabled=true \

/data/jars/SqlOnSpark.jar \

 "SELECT COUNT(*) FROM xx"