Issue jps and get the process id or Try to get the process id of datanode. Issue ps-fu userid of the user through which datanode is running.
Then kill the process using kill -9 On 28 Feb 2015 09:38, "Daniel Klinger" <[email protected]> wrote: > Hello, > > > > I used a lot of Hadoop-Distributions. Now I’m trying to install a pure > Hadoop on a little „cluster“ for testing (2 CentOS-VMs: 1 Name+DataNode 1 > DataNode). I followed the instructions on the Documentation site: > http://hadoop.apache.org/docs/current/hadoop-project-dist/hadoop-common/ClusterSetup.html > . > > > > I’m starting the Cluster like it is described in the Chapter „Operating > the Hadoop Cluster“(with different users). The starting process works > great. The PID-Files are created in /var/run and u can see that Folders and > Files are created in the Data- and NameNode folders. I’m getting no errors > in the log-files. > > > > When I try to stop the cluster all Services are stopped (NameNode, > ResourceManager etc.). But when I stop the DataNodes I’m getting the > message: „No DataNode to stop“. The PID-File and the in_use.lock-File are > still there and if I try to start the DataNode again I’m getting the error > that the Process is already running. When I stop the DataNode as hdfs > instead of root the PID and in_use-File are removed but I’m still getting > the message: „No DataNode to stop“ > > > > What I’m doing wrong? > > > > Greets > > dk >
