Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo >> mail # user >> cloud configuration

Copy link to this message
cloud configuration
Hello all,
I am working to setup a small cloud of 8 nodes.  All slave nodes are attached to a standalone switch.  The master node, which has zookeeper, namenode and jobtracker has two nics: one on the standalone switch and one to the main network.  
I was hoping that I could run java code to query the cloud through the master from any system on the main network.  Is this possible or will I have to run the code on the master node with this setup?  Currently I am getting this error:

-Test Java
ZooKeeperInstance zki = new ZooKeeperInstance("cloud", "master-server");
System.err.println("Got zookeeper instance!");
Connector conn = zki.getConnector("root", "somepassword");
System.err.println("never get here.");

12/03/26 23:24:03 INFO zookeeper.ZooKeeper: Initiating client connection, connectString=master-server sessionTimeout=30000 watcher=org.apache.accumulo.core.zookeeper.ZooSession$AccumuloWatcher@85af80
12/03/26 23:24:03 INFO zookeeper.ClientCnxn: Opening socket connection to server /  (fake IP)
12/03/26 23:24:03 INFO zookeeper.ClientCnxn: Socket connection established to master-server/, initiating session
12/03/26 23:24:03 INFO zookeeper.ClientCnxn: Session establishment complete on server master-server/, sessionid = 0x1365226ee5d0039, negotiated timeout = 30000
Got zookeeper instance!
12/03/26 23:24:25 WARN impl.ServerClient: Failed to find an available server in the list of servers: [ (120000), (120000)]

Thanks in advance,