Mailing-List: contact general-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: general@hadoop.apache.org
Received-SPF: pass (athena.apache.org: domain of marcoscba@gmail.com
 designates 74.125.82.176 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=mrAFsDCV6+FDT4UTGvrbEvHvZnaiiCv7NeK78B+jYEhbaEN4Z9NkAUku/SWZfujtub
         2eCrBf290iZfs64Q8xV8a/bwO6Wg2lcBudwqC3N/8O+cr6/oE4uCJv3+KErgymXSKV4e
         Z00TbZLADUBgcaE+sYz1TkpmbWknjrmiuz1pc=
MIME-Version: 1.0
In-Reply-To: <AANLkTik_dZctLp6Af1k9C75oJ+JhboRGZOXGsYOfwZLK@mail.gmail.com>
References: <AANLkTin6ZUrWYUscAXmq0bOG8mWA35xcAfvLD-jYVzwu@mail.gmail.com>
	<AANLkTik_dZctLp6Af1k9C75oJ+JhboRGZOXGsYOfwZLK@mail.gmail.com>
Date: Sat, 2 Oct 2010 17:16:08 -0400
Message-ID: <AANLkTi=+r_qR093O=O0+2C=rRfLsoJRz=x5AWuRKgHoF@mail.gmail.com>
Subject: Re: Is it possible write directly to datanode's?
From: Marcos Pinto <marcoscba@gmail.com>
To: general@hadoop.apache.org
Content-Type: multipart/alternative; boundary=001636d34693e0e26e0491a8d163

--001636d34693e0e26e0491a8d163
Content-Type: text/plain; charset=ISO-8859-1

Thanks Guys, I misunderstood the picture in the hadoop site. Now I know that
the namenode says in which datanode should be written and which part of
file. So I dont need to worry about it. Thanks.

On Sat, Oct 2, 2010 at 2:57 PM, Ryan Rawson <ryanobjc@gmail.com> wrote:

> I suggest you read this:
> http://hadoop.apache.org/common/docs/r0.20.0/hdfs_design.html
>
> -ryan
>
> On Fri, Oct 1, 2010 at 8:06 AM, Marcos Pinto <marcoscba@gmail.com> wrote:
> > Hi guys, I gotta a question and I really would appreciate some help,
> > suggestion, or advices.
> > Is it possible write directly to the datanode's and eliminate the
> bottleneck
> > to write into HDFS?
> > for example: I gotta 40 writers and I writes th data direct to namenode,
> i
> > was wondering if it would become a bottleneck so I would like to know
> > if it is possible to write the data direct to the datanode's.
> > I've read that Hadoop doesnt support multiple writes so how can we deal
> with
> > big cluster? I mean I am really going to have this bottleneck???
> >
> > Thanks in advanced.
> >
>

--001636d34693e0e26e0491a8d163--