Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HDFS, mail # user - Re: HDFS interfaces


Copy link to this message
-
RE: HDFS interfaces
John Lilley 2013-06-04, 18:35
When you use the HDFS client interface to read a file, it automatically figures out which datanodes to contact for reading which blocks.  There isn't really a "main" block.  However I have read that the first location listed for each block is the "recommended" one to read for an outside client.  Normally, an outside client doesn't need to know this information at all as the HDFS file interface takes care of it.  An "inside" application such as MapReduce *does* need to know this information so that it can run tasks on nodes that are "close" to the data split being processed.  If you are writing a custom ApplicationMaster using YARN, you will also want to know this.

John

From: Mahmood Naderan [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, June 04, 2013 12:01 AM
To: [EMAIL PROTECTED]
Subject: Re: HDFS interfaces

There are many instances of getFileBlockLocations in hadoop/fs. Can you explain which one is the main?
>It must be combined with a method of logically splitting the input data along block boundaries, and of launching tasks on worker nodes that >are close to the data splits
Is this a user level task of system level task?

Regards,
Mahmood

________________________________
From: John Lilley <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>>
To: "[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>" <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>>; Mahmood Naderan <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>>
Sent: Tuesday, June 4, 2013 3:28 AM
Subject: RE: HDFS interfaces

Mahmood,

It is the in the FileSystem interface.
http://hadoop.apache.org/docs/current/api/org/apache/hadoop/fs/FileSystem.html#getFileBlockLocations(org.apache.hadoop.fs.Path, long, long)<http://hadoop.apache.org/docs/current/api/org/apache/hadoop/fs/FileSystem.html#getFileBlockLocations(org.apache.hadoop.fs.Path,%20long,%20long)>

This by itself is not sufficient for application programmers to make good use of data locality.  It must be combined with a method of logically splitting the input data along block boundaries, and of launching tasks on worker nodes that are close to the data splits.  MapReduce does both of these things internally along with the file-format input classes.  For an application to do so directly, see the new YARN-based interfaces ApplicationMaster and ResourceManager.  These are however very new and there is little documentation or examples.

john

From: Mahmood Naderan [mailto:[EMAIL PROTECTED]]
Sent: Monday, June 03, 2013 12:09 PM
To: [EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>
Subject: HDFS interfaces

Hello,
It is stated in the "HDFS architecture guide" (https://hadoop.apache.org/docs/r1.0.4/hdfs_design.html) that

HDFS provides interfaces for applications to move themselves closer to where the data is located.

What are these interfaces and where they are in the source code? Is there any manual for the interfaces?

Regards,
Mahmood