I'm going through the big data mini course ( http://ampcamp.berkeley.edu/big-data-mini-course/launching-a-bdas-cluster-on-ec2.html) and am getting urllib2 connection refused errors. They're getting thrown in check_spark_cluster() function when waiting for the cluster to start. The URL is "http://<master-node>:8080/json".
A curl against the URL succeeds if done from either the master or the slaves, so it looks like an issue with the EC2 security groups. There's an entry for TCP port 8080 - 8081, IP 0.0.0.0/0 which should let the check succeed. I checked and there's no host firewall running. Anyone else seen this? Want to make sure it's not a known issue before I spend time debugging the network. Oh, it also looks like the URL's missing a trailing "/", without it the response is blank, with it I get the full JSON payload. Thanks, Dave
