DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:sender:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references:x-google-sender-auth;
        b=o8tYGHOABIN3x7XodnVda0nEP1kGtTl/SPyZCQjPXYHlSJljiHIL7S298DUTEpLVbZ9PgePK0zx/zagAiyAxhb2fukHBJtK9EmmSZXCyY33y04wGA0dLTsp6dqhznFgBST5wZGFazQjkAYxFotSrRnHSUYBkVNjFv5b9vX/nQ8M=
Message-ID: <c2213f680712170726m5fc2306arfca039b6c39af831@mail.gmail.com>
Date: Mon, 17 Dec 2007 10:26:50 -0500
From: "Alex Gorbachev" <ag@oracloid.com>
To: "Jason Arneil" <Jason@nominet.org.uk>
Subject: Re: does FRA is really needed ?
Cc: grzegorzof@interia.pl, oracle-l@freelists.org
In-Reply-To: <OFE306D386.A3581194-ON802573B4.004737F1-802573B4.0048E0D8@nominet.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
References: <47666775.9060106@interia.pl>
	 <OFE306D386.A3581194-ON802573B4.004737F1-802573B4.0048E0D8@nominet.org.uk>
Sender: oracle-l-bounce@freelists.org
Errors-to: oracle-l-bounce@freelists.org
Precedence: normal
Reply-to: ag@oracloid.com

On Dec 17, 2007 8:16 AM, Jason Arneil <Jason@nominet.org.uk> wrote:
> The only reason for the FRA in a dataguard environment is if you would
> like, when you perform a failover, to have the ability to flashback the
> (now) old primary and reinstantiate it as a standby (assuming the hardware
> is useable). If you do not have flashback database set up on your primary
> when you perform a failover you will basically have to recreate a standby
> on your old primary using a backup.

That's what I thought so the real requirement is to have flashback
enabled. FRA comes as a consequence.
Thanks for confirming that, Jason.

> I would think carefully, if you really want to go down the road of rolling
> your own, rather than relying on what you have already paid Oracle for.
> Why try and re-invent the wheel?

This is not re-inventing the wheel. Physical standby doesn't work for
data warehouses loaded with NOLOGGING.
There are two common ways I'm aware of to have DR for DW:
- have two parallel DW running with data load job running for both
- incrementally updated standby

I agree it's not straightforward Data Guard configuration and requires
much more efforts and careful consideration how it's maintained and
how switchover and failover scenarios are handled.

100 GB/hour is actually a lot and I'm not sure if OP meant that this
is current redo generation rate with bulk of loads done with NOLOGGING
or it's estimated rate with FORCE LOGGING.

Cheers,
Alex

-- 
Alex Gorbachev, Oracle DBA Brewer, The Pythian Group
http://www.pythian.com/blogs/author/alex http://www.oracloid.com
BAAG party - www.BattleAgainstAnyGuess.com
--
http://www.freelists.org/webpage/oracle-l