You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@hadoop.apache.org by Clement Jebakumar <je...@gmail.com> on 2010/11/05 02:47:57 UTC

Keeping Different folders for Different Hadoop Datanodes

Hello,

Is there a way to specify different folder paths to different datanode
instances in Hadoop?
I need to attach a another harddisk to a PC and want to use it for hadoop.
Is it possible?

*Clement Jebakumar,*
111/27 Keelamutharamman Kovil Street,
Tenkasi, 627 811
http://www.declum.com/clement.html

Re: Keeping Different folders for Different Hadoop Datanodes

Posted by Harsh J <qw...@gmail.com>.
Hi,

On Fri, Nov 5, 2010 at 7:17 AM, Clement Jebakumar <je...@gmail.com> wrote:
> Hello,
>
> Is there a way to specify different folder paths to different datanode
> instances in Hadoop?

The property "dfs.data.dir" can be a comma separated list of paths,
enabling a DataNode to use more than one mount point for itself.

To run multiple DataNode instances itself on a single machine (which
isn't a good idea in production), see this discussion:
http://search-hadoop.com/m/sApJY1zWgQV

> I need to attach a another harddisk to a PC and want to use it for hadoop.
> Is it possible?

Yes, just modify the "dfs.data.dir" property to reflect the new mount
point location and restart your DataNode. Copy a selection of data
already residing on your HDFS, to balance things between the disks
(you can delete the older copy of those files after the operation).

>
> *Clement Jebakumar,*
> 111/27 Keelamutharamman Kovil Street,
> Tenkasi, 627 811
> http://www.declum.com/clement.html
>



-- 
Harsh J
www.harshj.com