On Wed, Mar 28, 2012 at 9:53 PM, Bing Li <
lblabs@...> wrote:
> Dear Peter,
>
> When I just started the Ubuntu machine, there was nothing in /tmp.
>
> After starting $HADOOP/bin/start-dfs.sh and $HBase/bin/start-hbase.sh, the
> following files were under /tmp. Do you think anything wrong? Thanks!
>
> libing@greatfreeweb:/tmp$ ls -alrt
> total 112
> drwxr-xr-x 22 root root 4096 2012-03-28 14:17 ..
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48
> hadoop-libing-namenode.pid
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48
> hadoop-libing-datanode.pid
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48
> hadoop-libing-secondarynamenode.pid
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48
> hbase-libing-zookeeper.pid
> drwxr-xr-x 3 libing libing 4096 2012-03-29 04:48 hbase-libing
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48 hbase-libing-master.pid
> -rw-r--r-- 1 libing libing 5 2012-03-29 04:48
> hbase-libing-regionserver.pid
> drwxr-xr-x 2 libing libing 4096 2012-03-29 04:48 hsperfdata_libing
> drwxrwxrwt 4 root root 4096 2012-03-29 04:48 .
> -rw-r--r-- 1 libing libing 71819 2012-03-29 04:48
> jffi5395899026867792565.tmp
> libing@greatfreeweb:/tmp$
>
> Best,
> Bing
Hmmm, all these files are owned by user 'libing' ...
that is different from my set-up.
Which manual are you exactly following for the pseudo-distributed
installation? In the Cloudera manual that I followed (cdh3u2) there
was a mention of making different users IIRC.
Also, in my set-up the hadoop is started automatically at boot-up
from the scripts in
/etc/rc2.d/S20hadoop-...
where user root then performs an su to user
=> hdfs for the Name, Secondaryname, Data node
=> mapred for the Jobtracker and Tasktracker.
I am not sure it is actually the intention that you start the
5 hadoop processes with the
"... $HADOOP/bin/start-dfs.sh ..."
command as you describe.
I stop and start them with
sudo /etc/init.d/hadoop-0.20-datanode {stop|start}
sudo /etc/init.d/hadoop-0.20-namenode {stop|start}
sudo /etc/init.d/hadoop-0.20-secondarynamenode {stop|start}
sudo /etc/init.d/hadoop-0.20-tasktracker {stop|start}
sudo /etc/init.d/hadoop-0.20-jobtracker {stop|start}
and this seems to be stable for now :-)
But maybe the manual that you follow gives other advise ?
HTH (not sure, I am a beginner too ...)
Peter