Mapreduce returns error when accessing to datanode on master machine - hadoop

Question

I set up a Hadoop 2.4.0 cluster with three machines. One master machine is deployed with namenode, resource manager, datanode and node manager. The other two worker machines are deployed with datanode and node manager. When I run Hive query, the work fails and the error is
2014-06-11 13:40:13,364 WARN [main] org.apache.hadoop.mapred.YarnChild: Exception running child : java.net.ConnectException: Call From master/127.0.0.1 to
master:43607 failed on connection exception: java.net.ConnectException: Connection >refused; For more details see: http://wiki.apache.org/hadoop/ConnectionRefused
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:5>7)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImp>l.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:526)
at org.apache.hadoop.net.NetUtils.wrapWithMessage(NetUtils.java:783)
at org.apache.hadoop.net.NetUtils.wrapException(NetUtils.java:730)
at org.apache.hadoop.ipc.Client.call(Client.java:1414)
at org.apache.hadoop.ipc.Client.call(Client.java:1363)
at org.apache.hadoop.ipc.WritableRpcEngine$Invoker.invoke(WritableRpcEngine.java:231)
at com.sun.proxy.$Proxy9.getTask(Unknown Source)
at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:136)
Caused by: java.net.ConnectException: Connection refused
at sun.nio.ch.SocketChannelImpl.checkConnect(Native Method)
at sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:739)
at org.apache.hadoop.net.SocketIOWithTimeout.connect(SocketIOWithTimeout.java:206)
at org.apache.hadoop.net.NetUtils.connect(NetUtils.java:529)
at org.apache.hadoop.net.NetUtils.connect(NetUtils.java:493)
at org.apache.hadoop.ipc.Client$Connection.setupConnection(Client.java:604)
at org.apache.hadoop.ipc.Client$Connection.setupIOstreams(Client.java:699)
at org.apache.hadoop.ipc.Client$Connection.access$2800(Client.java:367)
at org.apache.hadoop.ipc.Client.getConnection(Client.java:1462)
at org.apache.hadoop.ipc.Client.call(Client.java:1381)
... 4 more
if I disable the datanode on master machine, everything works well. I'm wondering if it's allowed to deployed datanode on the master machine. Thank you for your kindly help in advance.
BTW, my /etc/hosts on the three machines are the same:
127.0.0.1 localhost
10.1.154.231 master
10.1.153.220 slave1
10.1.153.133 slave2

Answer 1

Please set up passwordless ssh on your master to itself.
You can achieve this by
cat ~/id_rsa.pub >> ~/.ssh/authorized_keys2
Make sure the permissions are correct
chmod 0600 ~/.ssh/authorized_keys2

Answer 2

In this case you may check if the namenode is started correctly on the master by checking logs at your yourhadoopfolder/logs/hadoop-[hadoop-user]-namenode-master.log
It is often caused by the hdfs is not formatted before. Run
hadoop namenode -format
Of course you will need to put your data to the cluster again.