Maybe the user who issues stop-all.sh does not have permissions to terminate the process of NN (and some other, depending on who/what started it). Check jps listing after stopping and with some ps/top checks, switch to the proper user and issue a stop-all again?
You can also issue it a SIGTERM I believe. On Mon, Aug 30, 2010 at 12:22 PM, vaibhav negi <[email protected]> wrote: > Hi , > > I am running hadoop 0.20.2 . with 2 node cluster. > I executed script stop-all.sh . But still 2 line logs are getting created > every hour in log directory of name node log directory. > How to completely shutdown hadoop cluster. > Below is the 1 line log. > > > 2010-08-29 00:30:00,018 INFO org.apache.hadoop.hdfs.server. > namenode.FSNamesystem.audit: ugi=root,root,bin,daemon,sys,adm,disk,wheel > ip=/10.0.8.47 cmd=listStatus src=/user dst=null > perm=null > > > Vaibhav Negi > -- Harsh J www.harshj.com
