I simply use the /sbin/hadoop-daemon.sh {start|stop} {service} script to control daemons at my end.
Does this not work for you? Or perhaps this thread is more about documenting that? 2011/11/30 Nitin Khandelwal <nitin.khandel...@germinait.com>: > Hi, > > Even i am facing the same problem. There may be some issue with script . > The doc says to start namenode type : > bin/hdfs namenode start > > But "start" is not recognized. There is a hack to start namenode with > command "bin/hdfs namenode &" , but no idea how to stop. > If it had been a issue with config , the later also should not have worked. > > Thanks, > Nitin > > > 2011/11/30 cat fa <boost.subscrib...@gmail.com> > >> In fact it's me to say sorry. I used the word "install" which was >> misleading. >> >> In fact I downloaded a tar file and extracted it to /usr/bin/hadoop >> >> Could you please tell me where to point those variables? >> >> 2011/11/30, Prashant Sharma <prashant.ii...@gmail.com>: >> > I am sorry, I had no idea you have done a rpm install, my suggestion was >> > based on the assumption that you have done a tar extract install where >> all >> > three distribution have to extracted and then export variables. >> > Also I have no experience with rpm based installs - so no comments about >> > what went wrong in your case. >> > >> > Basically from the error i can say that it is not able to find the jars >> > needed on classpath which is referred by scripts through >> > HADOOP_COMMON_HOME. I would say check with the access permission as in >> > which user was it installed with and which user is it running with ? >> > >> > On Tue, Nov 29, 2011 at 10:48 PM, cat fa <boost.subscrib...@gmail.com >> >wrote: >> > >> >> Thank you for your help, but I'm still a little confused. >> >> Suppose I installed hadoop in /usr/bin/hadoop/ .Should I >> >> point HADOOP_COMMON_HOME to /usr/bin/hadoop ? Where should I >> >> point HADOOP_HDFS_HOME? Also to /usr/bin/hadoop/ ? >> >> >> >> 2011/11/30 Prashant Sharma <prashant.ii...@gmail.com> >> >> >> >> > I mean, you have to export the variables >> >> > >> >> > export HADOOP_CONF_DIR=/path/to/your/configdirectory. >> >> > >> >> > also export HADOOP_HDFS_HOME ,HADOOP_COMMON_HOME. before your run your >> >> > command. I suppose this should fix the problem. >> >> > -P >> >> > >> >> > On Tue, Nov 29, 2011 at 6:23 PM, cat fa <boost.subscrib...@gmail.com> >> >> > wrote: >> >> > >> >> > > it didn't work. It gave me the Usage information. >> >> > > >> >> > > 2011/11/29 hailong.yang1115 <hailong.yang1...@gmail.com> >> >> > > >> >> > > > Try $HADOOP_PREFIX_HOME/bin/hdfs namenode stop --config >> >> > $HADOOP_CONF_DIR >> >> > > > and $HADOOP_PREFIX_HOME/bin/hdfs datanode stop --config >> >> > $HADOOP_CONF_DIR. >> >> > > > It would stop namenode and datanode separately. >> >> > > > The HADOOP_CONF_DIR is the directory where you store your >> >> configuration >> >> > > > files. >> >> > > > Hailong >> >> > > > >> >> > > > >> >> > > > >> >> > > > >> >> > > > *********************************************** >> >> > > > * Hailong Yang, PhD. Candidate >> >> > > > * Sino-German Joint Software Institute, >> >> > > > * School of Computer Science&Engineering, Beihang University >> >> > > > * Phone: (86-010)82315908 >> >> > > > * Email: hailong.yang1...@gmail.com >> >> > > > * Address: G413, New Main Building in Beihang University, >> >> > > > * No.37 XueYuan Road,HaiDian District, >> >> > > > * Beijing,P.R.China,100191 >> >> > > > *********************************************** >> >> > > > >> >> > > > From: cat fa >> >> > > > Date: 2011-11-29 20:22 >> >> > > > To: common-user >> >> > > > Subject: Re: [help]how to stop HDFS >> >> > > > use $HADOOP_CONF or $HADOOP_CONF_DIR ? I'm using hadoop 0.23. >> >> > > > >> >> > > > you mean which class? the class of hadoop or of java? >> >> > > > >> >> > > > 2011/11/29 Prashant Sharma <prashant.ii...@gmail.com> >> >> > > > >> >> > > > > Try making $HADOOP_CONF point to right classpath including your >> >> > > > > configuration folder. >> >> > > > > >> >> > > > > >> >> > > > > On Tue, Nov 29, 2011 at 3:58 PM, cat fa < >> >> boost.subscrib...@gmail.com >> >> > > >> >> > > > > wrote: >> >> > > > > >> >> > > > > > I used the command : >> >> > > > > > >> >> > > > > > $HADOOP_PREFIX_HOME/bin/hdfs start namenode --config >> >> > $HADOOP_CONF_DIR >> >> > > > > > >> >> > > > > > to sart HDFS. >> >> > > > > > >> >> > > > > > This command is in Hadoop document (here >> >> > > > > > < >> >> > > > > > >> >> > > > > >> >> > > > >> >> > > >> >> > >> >> >> http://hadoop.apache.org/common/docs/r0.23.0/hadoop-yarn/hadoop-yarn-site/ClusterSetup.html >> >> > > > > > >) >> >> > > > > > >> >> > > > > > However, I got errors as >> >> > > > > > >> >> > > > > > Exception in thread "main" >> java.lang.NoClassDefFoundError:start >> >> > > > > > >> >> > > > > > Could anyone tell me how to start and stop HDFS? >> >> > > > > > >> >> > > > > > By the way, how to set Gmail so that it doesn't top post my >> >> reply? >> >> > > > > > >> >> > > > > >> >> > > > >> >> > > >> >> > >> >> >> > >> > > > > -- > > > Nitin Khandelwal > -- Harsh J