You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-dev@hadoop.apache.org by Michael Jones <mc...@gmail.com> on 2018/01/04 21:53:11 UTC
The command "hdfs dfsadmin -report" fails because of a "connection
exception: java.net.ConnectException:
I have two Ubuntu 16.04 servers running open source Hadoop 3.0. One is a
NameNode, and the other is a DataNode. I have configured three .xml files
on both servers to establish a cluster. The hduser on the NameNode can ssh
into the DataNode. I have configured a workers file on the NameNode.
On the NameNode server I run this: bash start-all.sh
On the DataNode I used the jps command before and after running the above
bash script on the NameNode. This bash script on the NameNode starts the
services on the datanode. I can tell because of the jps command's output.
On the NameNode I run this: "hdfs dfsadmin -report" but I get this:
"report: Call From NameNodeServer/10.10.10.10 to NameNodeServer:54310
failed on connection exception: java.net.ConnectException: Connection
refused; For more details see:
http://wiki.apache.org/hadoop/ConnectionRefused"
The /etc/hosts files of both servers have no entry for 127.0.0.1 nor one
for localhost.
My .xml files refer to the correct NameNode server. I could manually ping
it from both servers. There is no firewall between the servers blocking
connectivity (as evidenced by start-all.sh successfully starting services).
The problem does not happen with two RedHat servers in the cluster.
What am I doing wrong with the two Ubuntu servers? Is this a bug?