I am attempting to submit a job (using SparkPi) from one Linux machine (Ubuntu 14.04) to Spark 1.3.1 running in standalone mode on another Linux machine (Xubuntu 12.04; spartacus.servile.war), but I cannot make a connection.
I have investigated everything I can think of to diagnose/fix the problem but have run out of ideas. Here are the facts; On the Xubuntu machine I can submit SparkPi without a problem. I can also test successfully that the master is listening on port 7077 by connecting with Telnet. Netstat shows: tcp6 0 0 spartacus.servile.war:7077 [::]:* LISTEN Iptables is not running, it is not even installed. I have log4j set to log in DEBUG mode to a file. On the Ubuntu client machine I can view the Spark Master web page at port 8080: http://spartacus:8080/ I can of course telnet to port 8080 on spartacus as well. If I try to telnet to port 7077 I get "connection refused". If I try to submit SparkPI on this machine like so: ./bin/spark-submit --class org.apache.spark.examples.SparkPi --master spark://spartacus.servile.war:7077 --executor-memory 10G --total-executor-cores 8 /home/carey/dev/spark-1.3.1-bin-hadoop2.6/lib/spark-examples-1.3.1-hadoop2.6.0.jar 1 I get the following messages: 15/05/20 13:38:19 WARN AppClient$ClientActor: Could not connect to akka.tcp://sparkmas...@spartacus.servile.war:7077: akka.remote.InvalidAssociation: Invalid address: akka.tcp://sparkmas...@spartacus.servile.war:7077 15/05/20 13:38:19 WARN Remoting: Tried to associate with unreachable remote address [akka.tcp://sparkmas...@spartacus.servile.war:7077]. Address is now gated for 5000 ms, all messages to this address will be delivered to dead letters. Reason: Connection refused: spartacus.servile.war/ 192.168.0.113:7077 Using "spartacus" or "192.168.0.113" instead of "spartacus.servile.war" makes no difference. Absolutely nothing shows up in the Spark log on spartacus when I try to submit, I just see the worker heartbeat exchange. In my hosts file on this machine I have: 192.168.0.113 spartacus.servile.war spartacus Using the default spark-env.sh or setting: export SPARK_MASTER_IP=spartacus.servile.war (or just spartacus, or 192.168.0.113) makes no difference. I have tried each combination of host ID in the submit and in the spark-env.sh file together (3x4 = 12 combinations) with the same result each time. Iptables is not running on the Ubuntu machine either. What is it I am missing?
