1. Create File System
FileSystem is an abstract class extending Hadoops configured class.
Like a Normal file system it defines all File functions.
All File System that Hadoop use indirectly extends FileSystem class.
Which File System to be invoked depends on Hadoop Configuration file.
Suppose we specify hdfs://localhost:9000 which tells hadoop uses DistributedFileSystem.
This file can be invoked from core-default.xml
file:/// invokes LocalFileSystem
2.Client invokes Server
mkdirs function assume calls the DistributedFileSystems mkdir
public boolean mkdirs(Path f, FsPermission permission) throws IOException {
return mkdirsInternal(f, permission, true);
}
mkdir() calls mkdirInternal which inturn uses variable : dfs which is object of DFSClient
Apache File Descritption
DFSClient can connect to a Hadoop Filesystem and perform basic file tasks. It uses the ClientProtocol to communicate with a NameNode daemon, and connects directly to DataNodes to read / write block data. Hadoop DFS users should obtain an instance of DistributedFileSystem, which uses DFSClient to handle filesystem tasks.
Initialize Method is called.
All operation on DistributedFileSystem transferred to DFSClient.
public DFSClient(URI nameNodeUri, ClientProtocol rpcNamenode,
Configuration conf, FileSystem.Statistics stats)
throws IOException {
DFSclient method call forwarded to namenode and its assoicate variable rpcNameNode
Client Protocol provide all method through which methods from client to server can be invoked.
All method calls from rpcNamenode come down to Invokers invoke() method.
This way the calls are transferred to server side method calls.
Eventually it gets connected to Server class.
3 major components of server
Listener, Responder, Handler,
...
