In HDFS file will will be devided based on dfs.blocksize configuration. You can see the code at DFSClient where the client will write the block of data. You can post it to mapreduce-...@hadoop.apache.org to get moredetails from mapreduce as well. In mapreduce, tasktrackers will run on each DN machines. Job tracker will assign the work to them. They will read the blocks and process.
I would suggest you to read Hadoop definitive guide to get more understanding of the system. Regards, Uma ________________________________________ From: Hadoop Sai [hadoop....@gmail.com] Sent: Thursday, November 24, 2011 10:28 PM To: common-dev@hadoop.apache.org Subject: help in learning hadoop How the file will be divided into blocks and how the map reduce read those blocks of data .Can anyone can point out this piece of code.
