Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
HDFS, mail # user - Namenode and Jobtracker dont start


+
Björn-Elmar Macek 2012-07-18, 14:29
+
Suresh Srinivas 2012-07-18, 17:47
+
Björn-Elmar Macek 2012-07-20, 14:15
+
Björn-Elmar Macek 2012-07-20, 14:54
+
Mohammad Tariq 2012-07-20, 14:58
+
Björn-Elmar Macek 2012-07-20, 15:38
Copy link to this message
-
Re: Namenode and Jobtracker dont start
Harsh J 2012-07-20, 16:01
Hi,

 <property>
        <name>dfs.hosts</name>
<value>/home/fb16/bmacek/hadoop-1.0.2/conf/hosts</value>
    </property>

This one is probably the cause of all your trouble. It makes the
"hosts" file a white-list of allowed nodes. Ensure, hence, that
"its-cs103.its.uni-kassel.de" is in this file for sure.

Also, dfs.hosts must be in hdfs-site.xml, and mapred.hosts in
mapred-site.xml, but you've got both of them in the latter. You should
fix this up as well.

Or if you do not need such a white-lister feature, just remove both
properties away and restart.

On Fri, Jul 20, 2012 at 9:08 PM, Björn-Elmar Macek
<[EMAIL PROTECTED]> wrote:
> Hi Mohammad,
>
> Thanks for your fast reply. Here they are:
>
> \_____________hadoop-env.sh___
> I added those 2 lines:
>
> # The java implementation to use.  Required.
> export JAVA_HOME=/opt/jdk1.6.0_01/
> export JAVA_OPTS="-Djava.net.preferIPv4Stack=true $JAVA_OPTS"
>
>
> \_____________core-site.xml_____
> <?xml version="1.0"?>
> <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
>
> <!-- Put site-specific property overrides in this file. -->
>
> <configuration>
>     <property>
>         <name>fs.default.name</name>
>         <value>hdfs://its-cs100:9005</value>
>     </property>
> </configuration>
>
>
> \_____________hdfs-site.xml____
> <?xml version="1.0"?>
> <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
>
> <!-- configure data paths for masters and slaves -->
>
> <configuration>
>     <property>
>         <name>dfs.name.dir</name>
>         <value>/home/work/bmacek/hadoop/master</value>
>     </property>
>     <!-- maybe one cannot config masters and slaves on with the same file
> -->
>     <property>
>         <name>dfs.data.dir</name>
> <value>/home/work/bmacek/hadoop/hdfs/slave</value>
>     </property>
>     <property>
>         <name>hadoop.tmp.dir</name>
> <value>/home/work/bmacek/hadoop/hdfs/tmp</value>
>     </property>
>
>     <property>
>         <name>dfs.replication</name>
>         <value>1</value>
>     </property>
> </configuration>
>
>
> \_______mapred-site.xml____
> <?xml version="1.0"?>
> <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
>
> <!-- Put site-specific property overrides in this file. -->
>
> <configuration>
>     <!-- master -->
>     <property>
>         <name>mapred.job.tracker</name>
>         <value>its-cs100:9004</value>
>     </property>
>     <!-- datanode -->
>     <property>
>         <name>dfs.hosts</name>
> <value>/home/fb16/bmacek/hadoop-1.0.2/conf/hosts</value>
>     </property>
>
>     <property>
>         <name>mapred.hosts</name>
> <value>/home/fb16/bmacek/hadoop-1.0.2/conf/hosts</value>
>     </property>
> </configuration>
>
> \_______masters____
> its-cs101
>
> \_______slaves______
> its-cs102
> its-cs103
>
>
> Thats about it, i think. I hope i didnt forget anything.
>
> Regards,
> Björn-Elmar
>
> Am 20.07.2012 16:58, schrieb Mohammad Tariq:
>
>> Hello sir,
>>
>>        If possible, could you please paste your config files??
>>
>> Regards,
>>      Mohammad Tariq
>>
>>
>> On Fri, Jul 20, 2012 at 8:24 PM, Björn-Elmar Macek
>> <[EMAIL PROTECTED]> wrote:
>>>
>>> Hi together,
>>>
>>> well just stumbled upon this post:
>>>
>>> http://ankitasblogger.blogspot.de/2012/01/error-that-occured-in-hadoop-and-its.html
>>>
>>> And it says:
>>> "Problem: Hadoop-datanode job failed or datanode not running:
>>> java.io.IOException: File ../mapred/system/jobtracker.info could only be
>>> replicated to 0 nodes, instead of 1.
>>> ...
>>> Cause: You may also get this message due to permissions. May be
>>> JobTracker
>>> can not create jobtracker.info on startup."
>>>
>>> Since the file does not exist i think, this might be a probable reason
>>> for
>>> my errors. But why should the JobTracker not be able to create that file.
>>> It
>>> created several other directories on this node with easy via the slave.sh
>>> script that i started with the very same user that calls start-all.sh.
>>>
>>> Any help would be really appreciated.

Harsh J
+
Mohammad Tariq 2012-07-20, 15:44