spark-submit任务提交

詹学伟
詹学伟
发布于 2024-04-21 / 6 阅读
0
0

spark-submit任务提交

Run application locally(本地运行)

/usr/local/spark-3.3.0/bin/spark-submit \
--master local[8] \
--class com.zxw.spark.Etl \
/opt/zxw-spark-1.0.jar hdfs://node3:8020/sdk/20230418/logs-.* hdfs://node3:8020/data/locally

Run on a Spark standalone cluster in client deploy mode

/usr/local/spark-3.3.0/bin/spark-submit \
--master spark://node3:7077 \
--class com.zxw.spark.Etl \
--driver-memory 1g \
--executor-memory 1g \
--total-executor-cores 8 \
/opt/zxw-spark-1.0.jar hdfs://node3:8020/sdk/20230418/logs-.* hdfs://node3:8020/data/spark-standalone

Run on a Spark standalone cluster in cluster deploy mode with supervise

/usr/local/spark-3.3.0/bin/spark-submit \
--master spark://node3:7077 \
--deploy-mode cluster \
--supervise \
--class com.zxw.spark.Etl \
--driver-memory 1g \
--executor-memory 1g \
--total-executor-cores 8 \
/opt/zxw-spark-1.0.jar hdfs://node3:8020/sdk/20230418/logs-.* hdfs://node3:8020/data/spark-cluster

Run on a YARN cluster export HADOOP_CONF_DIR=XXX

/usr/local/spark-3.3.0/bin/spark-submit \
--class com.zxw.spark.Etl \
--master yarn \
--driver-memory 8g \
--executor-memory 1g \
--total-executor-cores 6 \
--num-executors 2 \
--conf spark.shuffle.consolidateFiles=true \
--conf spark.driver.maxResultSize=2048M \
/opt/zxw-spark-1.0.jar hdfs://node3:8020/sdk/20230418/logs-.* hdfs://node3:8020/data/yarn

参考连接:https://blog.csdn.net/kezhong_wxl/article/details/77507361


评论