Mailing-List: contact common-user-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: common-user@hadoop.apache.org
Received-SPF: neutral (nike.apache.org: local policy)
Date: Wed, 14 Dec 2011 13:48:58 -0800
From: Konstantin Boudnik <cos@apache.org>
To: common-user@hadoop.apache.org
Subject: Re: HDFS Backup nodes
Message-ID: <20111214214858.GC27452@linspire.com>
References: <20111214072839.GB27452@linspire.com>
 <CB0E2490.645F9%scott@richrelevance.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <CB0E2490.645F9%scott@richrelevance.com>
User-Agent: Mutt/1.5.18 (2008-05-17)

On Wed, Dec 14, 2011 at 10:09AM, Scott Carey wrote:
> 
> On 12/13/11 11:28 PM, "Konstantin Boudnik" <cos@apache.org> wrote:
> 
> >On Tue, Dec 13, 2011 at 11:00PM, M. C. Srivas wrote:
> >> Suresh,
> >> 
> >> As of today, there is no option except to use NFS.  And as you yourself
> >> mention, the first HA prototype when it comes out will require NFS.
> >
> >
> >NFS is just happen to be readily available in any data center and doesn't
> >require much of the extra investment on top of what exists.
> 
> That is a false assumption.  I'm not buying a netapp filer just for this.
>  We have no NFS, or want any.  If we ever use it, it won't be in the data
> center with Hadoop!

It isn't a false assumption, it is a reasonable one based on the experience.
You don't need netapp for NFS, you can have a Thumper or whatever. I am not
saying NFS is the only and the best - all I said it is pretty common ;) I
would opt fo BK or Jini Spaces like solution any day, though.

Cos