Reputation: 13
does the number of input splits determine the number of running nodes in a mapreduce job.
i mean if i have 5 input splits will this make them distributed on 5 nodes to be processed. And is the number of map tasks is similar to number of nodes ?
How to check that in clouderavm
Upvotes: 0
Views: 612
Reputation: 1766
No splits do not determine number of nodes.
Hadoop Cluster i.e NameNode ,DataNodes and other service are setup by Hadoop Admin based on size of data.
On Cloudera VM open browser(Mozilla for CDH3) there is a bookmark Namenode status. Click on that and Namenode contains all the metadata about the cluster you are working on.
To know about job information executed by Map Reduce framework, Job Tracker is there.
Upvotes: 1