hdfs: Distributed File System
Hdfs common features with the file system:
- There directory structure, the top-level directory is: /
- The system is stored in file
- The system can provide file: create, delete, modify, view, movement and other functions
hdfs a difference with the ordinary stand-alone file system:
- Single file stored in the file system, is the operating system of a machine
- hdfs file system across many machines as N
- Single file stored in the file system is on a machine disk
- Hdfs file stored in the file system, a multi-machine n falls single local file system (hdfs linux is a local file system on the file system)
hdfs working mechanism:
- Customers put a file into hdfs, in fact, after hdfs will cut this document, distributed storage in N linux machine system (role is responsible for the storage of file blocks: the Data the Node ) <precisely: diced behavior by the customer the decision to end >
- Once the file is stored into cubes, then, hdfs in there must be a mechanism to record the user's cut information for each file, and each piece of machine-specific storage (responsible role block information is recorded: name the Node )
-
To ensure data security, hdfs can block each file in a cluster to store multiple copies ( in the end keep a few copies, it is specified by the time the file is stored in the client)
Summary: hdfs a system, a server running a namenode, N, and running of the servers datanode!
Working mechanism Figure