Mailing-List: contact user-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hbase.apache.org
Received-SPF: neutral (nike.apache.org: local policy)
MIME-Version: 1.0
In-Reply-To: 
 <2D6136772A13B84E95DF6DA79E85A9F001584353F1EA@NSPEXMBX-A.the-lab.llnl.gov>
References: <1310732977.2089.10.camel@localhost.localdomain>
 <2D6136772A13B84E95DF6DA79E85A9F001584353F1EA@NSPEXMBX-A.the-lab.llnl.gov>
From: Ted Dunning <tdunning@maprtech.com>
Date: Fri, 15 Jul 2011 11:33:05 -0700
Message-ID: 
 <CAND0qzv_GGd7-CHUV8WGAOWySPzPQ7Xe4e528oop1PgfOXK0qA@mail.gmail.com>
Subject: Re: Backup data for a single hbase instance
To: user@hbase.apache.org
Content-Type: multipart/alternative; boundary=20cf30780bfc8d99e204a81fe21e

--20cf30780bfc8d99e204a81fe21e
Content-Type: text/plain; charset=ISO-8859-1

This is basically the same process that I was referring to earlier except
that if you have the ability to take snapshots, you can do this without
downtime or using extra disk space for the copy.

To HBase's credit, you can do this while clients are doing a massive amount
of updates.  In fact, this provides an interesting way to test crash
resistance of HBase.

On Fri, Jul 15, 2011 at 9:52 AM, Buttler, David <buttler1@llnl.gov> wrote:

> Yes, I have done this in the past on a full cluster.  The backup process is
> slow for large data sets, but at least it is simple.
> Dave
>
>
> -----Original Message-----
> From: Joel Halbert [mailto:joel@su3analytics.com]
> Sent: Friday, July 15, 2011 5:30 AM
> To: users-hbase
> Subject: Backup data for a single hbase instance
>
> Hi,
>
> I  have a standalone, single, HBase instance (not running on Hadoop, or
> in a cluster).
>
> What's the best strategy for creating a backup of the data?
>
> I can afford downtime so is it safe to simply create a copy of the
> hbase.rootdir ?
>
> Thanks
> Joel
>
>

--20cf30780bfc8d99e204a81fe21e--