跳到主要内容
版本:2.1.0

deployment

Deployment and run

seatunnel v2 For Spark relies on the Java runtime environment and Spark. For detailed seatunnel installation steps, please refer to installing seatunnel

The following focuses on how different platforms operate:

Run seatunnel locally in local mode

./bin/start-seatunnel-spark.sh \
--master local[4] \
--deploy-mode client \
--config ./config/application.conf

Run seatunnel on Spark Standalone cluster

# client mode
./bin/start-seatunnel-spark.sh \
--master spark://ip:7077 \
--deploy-mode client \
--config ./config/application.conf

# cluster mode
./bin/start-seatunnel-spark.sh \
--master spark://ip:7077 \
--deploy-mode cluster \
--config ./config/application.conf

Run seatunnel on Yarn cluster

# client mode
./bin/start-seatunnel-spark.sh \
--master yarn \
--deploy-mode client \
--config ./config/application.conf

# cluster mode
./bin/start-seatunnel-spark.sh \
--master yarn \
--deploy-mode cluster \
--config ./config/application.conf

Run seatunnel on Mesos cluster

# cluster mode
./bin/start-seatunnel-spark.sh \
--master mesos://ip:7077 \
--deploy-mode cluster \
--config ./config/application.conf

For the meaning of the master and deploy-mode parameters of start-seatunnel-spark.sh , please refer to: Command Instructions

If you want to specify the resource size occupied by seatunnel when running, or other Spark parameters , you can specify it in the configuration file specified by --config :

env {
spark.executor.instances = 2
spark.executor.cores = 1
spark.executor.memory = "1g"
...
}
...

For how to configure seatunnel , please refer to seatunnel common configuration