How to handle dynamic port in Apache Spark / Hadoop Yarn-CodePudding

I have setup a Hadoop cluster with 1 name node and 2 data nodes. I've also installed Yarn and Spark on top of that in the name node.

I notice that whenever I try run the example jar here:

spark-submit --deploy-mode cluster --class org.apache.spark.examples.SparkPi $SPARK_HOME/examples/jars/spark-examples_*.jar 10

I will always get the no route to host exception:

Uncaught exception: org.apache.spark.SparkException: Exception thrown in awaitResult:
at org.apache.spark.util.ThreadUtils$.awaitResult(ThreadUtils.scala:301)
at org.apache.spark.rpc.RpcTimeout.awaitResult(RpcTimeout.scala:75)
at org.apache.spark.rpc.RpcEnv.setupEndpointRefByURI(RpcEnv.scala:102)
at org.apache.spark.rpc.RpcEnv.setupEndpointRef(RpcEnv.scala:110)
at org.apache.spark.deploy.yarn.ApplicationMaster.runExecutorLauncher(ApplicationMaster.scala:558)
at org.apache.spark.deploy.yarn.ApplicationMaster.run(ApplicationMaster.scala:277)
at org.apache.spark.deploy.yarn.ApplicationMaster$$anon$3.run(ApplicationMaster.scala:926)
at org.apache.spark.deploy.yarn.ApplicationMaster$$anon$3.run(ApplicationMaster.scala:925)
at java.base/java.security.AccessController.doPrivileged(Native Method)
at java.base/javax.security.auth.Subject.doAs(Subject.java:423)
at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1878)
at org.apache.spark.deploy.yarn.ApplicationMaster$.main(ApplicationMaster.scala:925)
at org.apache.spark.deploy.yarn.ExecutorLauncher$.main(ApplicationMaster.scala:957)
at org.apache.spark.deploy.yarn.ExecutorLauncher.main(ApplicationMaster.scala)
Caused by: java.io.IOException: Failed to connect to lnx-pen205/xx.xx.xx.xx:9222
at org.apache.spark.network.client.TransportClientFactory.createClient(TransportClientFactory.java:288)
at org.apache.spark.network.client.TransportClientFactory.createClient(TransportClientFactory.java:218)
at org.apache.spark.network.client.TransportClientFactory.createClient(TransportClientFactory.java:230)
at org.apache.spark.rpc.netty.NettyRpcEnv.createClient(NettyRpcEnv.scala:204)
at org.apache.spark.rpc.netty.Outbox$$anon$1.call(Outbox.scala:202)
at org.apache.spark.rpc.netty.Outbox$$anon$1.call(Outbox.scala:198)
at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
at java.base/java.lang.Thread.run(Thread.java:829)
Caused by: io.netty.channel.AbstractChannel$AnnotatedNoRouteToHostException: No route to host: lnx-pen205/xx.xx.xx.xx:9222
Caused by: java.net.NoRouteToHostException: No route to host
at java.base/sun.nio.ch.SocketChannelImpl.checkConnect(Native Method)
at java.base/sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:777)
at io.netty.channel.socket.nio.NioSocketChannel.doFinishConnect(NioSocketChannel.java:330)
at io.netty.channel.nio.AbstractNioChannel$AbstractNioUnsafe.finishConnect(AbstractNioChannel.java:334)
at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:710)
at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:658)
at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:584)
at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:496)
at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:986)
at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74)
at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30)
at java.base/java.lang.Thread.run(Thread.java:829)

I noticed that the port being used will be randomly assigned during the runtime, the example .jar will work if for example I set the spark.driver.port as 9222 then opening said port with the firewall. But then if any other session is started (for example, pyspark shell), it wouldn't start as the port is already in use.

My question is: How do I allow connections to the ports dynamically defined by Spark/Yarn? I read somewhere that I should disable the firewall, but that does not sound like a good idea.. Thanks in advance.

CodePudding user response：

There's spark.driver.port as well as spark.driver.blockManager.port. Both are starting ranges to spark.port.maxRetries (default 16).

So, you'll need to open at least 32 ports for these.

I did some testing with dynamic Spark ports in Mesos Docker a few years ago - https://stackoverflow.com/a/56486271/2308683