问题
I am trying to setup a standalone spark cluster on 2 machines within my organization network. Both are ubuntu 16.04 machines with same configuration.
There is passwordless ssh setup from master to slave and slave to master as well.
The config of the master and the slave node are mentioned below.
Master configuration:
Spark-version: 2.4.4
/etc/hosts : 10.x.x.2 master
10.x.x.3 slave01
127.0.0.1 dm.abc.net localhost dm
127.0.1.1
10.x.x.4 dm.abc.net dm
10.x.x.5 abc.net
/usr/local/spark/conf/slaves: 10.x.x.3
/usr/local/spark/conf/spark-env.sh: export SPARK_MASTER_HOST='10.x.x.2'
export SPARK_LOCAL_IP='10.x.x.2'
export JAVA_HOME='/usr/lib/jvm/java-8-openjdk-amd64/jre/'
Slave configuration:
Spark-version: 2.4.4
/etc/hosts : 10.x.x.2 master
10.x.x.3 slave01
127.0.0.1 dm1.abc.net localhost dm1
127.0.1.1
10.x.x.4 dm1.abc.net dm1
10.x.x.5 abc.net
/usr/local/spark/conf/slaves: 10.x.x.3
/usr/local/spark/conf/spark-env.sh: export SPARK_MASTER_HOST='10.x.x.2'
export SPARK_LOCAL_IP='10.x.x.3'
I try the following command on the master: /usr/local/spark/sbin/start-all.sh
But I keep getting: Failed to connect to master. error in the slave log. Any pointers.
来源：https://stackoverflow.com/questions/58767418/spark-slave-unable-to-connect-to-master