You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by Eric Palacios <ep...@gmail.com> on 2008/01/25 18:18:09 UTC

Problem with mapred.system.dir variable and local jobtracker

Hi all,

I recently added to the configuration this :

<property>
   <name>mapred.system.dir</name>
   <value>/hadoop/mapred/system</value>
 <description>The shared directory where MapReduce stores control files.
   </description>
</property>

in order to allow multiple users use the cluster, and works fine , but if i
want to run the mapreduce in local (local jobtracker) i get this exception :


java.io.IOException: Mkdirs failed to create
/hadoop/mapred/system/job_local_1

(though the /hadoop/mapred/system  folder is already existing in HDFS)

Do you have an idea  about why the jobs fail in this situation ?

Thanks in advance,

Eric.