All the slaves in the Hadoop cluster should be of the same configuration

Question

In Hadoop is it mandatory that all the slaves in the Hadoop cluster should be of the same configuration?

I suggest you look into using Apache Ambari for managing Hadoop and its configurations — OneCricketeer, Feb 12 '19 at 04:36

score 0 · Answer 1 · answered Feb 11 '19 at 09:38

0

Hadoop datanode can have different configuration like datanode total memory, datanode mount points etc.

Example: Datanode1 etc/hadoop/hdfs-site.xml can be like

<property>
    <name>dfs.datanode.data.dir</name>
    <value>/mount/data1,/mount/data2,/mount/data3</value>
</property>

Datanode2 etc/hadoop/hdfs-site.xml can be like

<property>
    <name>dfs.datanode.data.dir</name>
    <value>/mnt/dt1</value>
 </property>

answered Feb 11 '19 at 09:38

unwelcomed_user

I feel like the question was asking about the `slaves` file. – OneCricketeer Feb 12 '19 at 04:36

1 Answers1