Using Whirr 0.8.2 and setting up a single Cassandra node on EC2 the configuration is somehow wrong. I cannot connect to the Cassandra Thrift port 9160 on the client machine (from which Whirr created the cluster) even though it shows the proper IP (of the launching machine) in the security group allowed to access that port. I tried using both my local linux box and an EC2 instance as the launching machine with exactly the same issue.
Running netstat on the Cassandra node shows it is listening to localhost:9160. If I look at the Cassandra log it shows that, instead of using the properly generated /etc/cassandra/conf/cassandra.yaml which has rpc_address=0.0.0.0 and other proper settings it seems to be using /usr/local/apache-cassandra-1.2.16/conf/cassandra.yaml instead which has only the default settings which includes rpc_address=localhost. Opening a cqlsh also shows the cluster name as Test Cluster instead of Whirr. Thus far I've been unable to figure out why it does that, as the $CASSANDRA_CONF=/etc/cassandra/conf. My only supposition is that having the /usr/local/apache-cassandra-1.2.16/bin/../conf on the classpath is somehow overriding other settings for where to source the configuration. Michael Oczkowski, Ph.D. Big Data Architect & Chief Data Scientist Seeq Corporation 206-801-9339 x704 [cid:F25521F6-6CD3-434A-B057-562FA55E4AB6]<http://www.seeq.com/> Transforming industrial process data into actionable intelligence Connect: LinkedIn<http://www.linkedin.com/company/3126445?trk=tyah> | Facebook<https://www.facebook.com/pages/Seeq-Corporation/424546840977984> | Twitter<https://twitter.com/SeeqCorporation>
