Hadoop distributed file System-HDFS

Hadoop Distributed File System

HDFS is a management of hadoop file system for example every Operating System do have FS like in windows we have NTFS , FAT32 managing metadata about your directories and files same in hdfs Master node [namenode] is managing metadata about all file and directories that are present in whole cluster.

Hadoop is not a single tool and having distributed file System. when user upload data at HDFS then hadoop distribute data across multiple nodes. Hdfs and map-reduce are two base component of hadoop echo system.

Let us Understand the concept of hadoop distributed file system.
1- hadoop is working on cluster computing concept that is having master slave architecture.
2- Master and slaves are machine serving couple of services.
Master: 
    1-NameNode
    2-Job Tracker.
    3-Secondary namenode.

Slave:
    1-Data Nmode
    2-Task Tracker
    3- Child Jvm.

1-Job Tracker is master daemon and Task Tracker is slave daemon job tracker distribute job across all nodes where data partitions exist.

2-In every data node we have Task Tracker and child jvm all Task Tracker register themselves with
job tracker while they are working on hadoop job.

3-Job is executed by child jvm and all task tracker send report to the job tracker like ack. message.

4-All data nodes sends acknowledge message to the Master Namenode to report that they are still alive or they are working properly.

5- Secondary name node is working like a helper of namenode. the task of secondary namenode is to  update namenode with cluster current image file that is stored in fsimage .

6- So finally  Task Tracker send Report  to Job Tracker.Data Node send Report to  Name Node.
and Secondary Name Node update Master Name Node.