Veritas-bu

[Veritas-bu] Incremental RMAN backups on split BCV's

2003-01-21 18:42:53
Subject: [Veritas-bu] Incremental RMAN backups on split BCV's
From: Mark.Donaldson AT experianems DOT com (Donaldson, Mark)
Date: Tue, 21 Jan 2003 16:42:53 -0700
This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01C2C1A6.D0E22A50
Content-Type: text/plain

The bottleneck is at the tape itself.  We still have DLT7000-based libraries
and doing nightly filesystem backups of what may become 3TB of data will
take too long.  We're going to have to move to RMAN backups so we can start
doing incremental backups of the database files.  So the problem is not the
data source but the destination.

The attraction of doing RMAN backups from BCV's is two-fold.  One, I'm
off-host - there will be no load added to the production database servers
from the backup process.  Two, the BCV itself makes a good backup.  Should I
lose the DB while the BCV's are split, I can "reverse establish" (restore)
the BCV's over the primary copies, roll it forward, and be back up and
running without spinning tape.  

A full RMAN recovery of the DB from tape can be done by fully restoring the
tapes to the BCV's mounted on the media server, then synchronizing the BCV's
against the primary mirrors.  Still uncertain is if I can restore an RMAN
backup on my media server directly to the primary database server.  This
latter move would be necessary for a single-tablespace restoral rather than
a full DB restoral.

-M

-----Original Message-----
From: Darren Dunham [mailto:ddunham AT taos DOT com]
Sent: Tuesday, January 21, 2003 3:38 PM
To: Veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Incremental RMAN backups on split BCV's


> 
> The problem is that anticipated growth will make this difficult in the
> future - we won't be able to move enough data in the necessary
> timespan.

Can I ask what the bottleneck will be?  I would think that the
BCV/filesystem backup would have higher throughput than an RMAN backup.

> Therefore, we're looking at using Oracle RMAN to perform database
> incrementals.  However, I'm still wanting to perform the backup off-host
so
> the idea is to use RMAN to backup BCV's mounted on my media server.
> 
> The idea is this, and I'm looking for proponents or detractors to this
idea:
> 
> 1. split & import as above to our media server
> 2. Open the database on the media server
> 3. perform the RMAN full and/or incremental as necessary
> 4. shutdown the Oracle instance
> 5. incrementally re-establish.
> 6. rinse & repeat the next night
> 
> The thought is that since the instance on the media server is opened &
> closed normally and since the BCV should be a blockwise identical copy of
> the original database, then RMAN doesn't know from run-to-run that this is
a
> "new" copy of the DB.  As far as RMAN is concerned, the only DB involved
> lives exclusively on the media server - it doesn't have to know about the
> original mounted on the main server.

Sounds workable to me, but I'd want to test the restore.  Do you have to
have the other machine involved, or could you restore directly to the
main server?

-- 
Darren Dunham                                           ddunham AT taos DOT com
Unix System Administrator                    Taos - The SysAdmin Company
Got some Dr Pepper?                           San Francisco, CA bay area
         < This line left intentionally blank to confuse you. >
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

------_=_NextPart_001_01C2C1A6.D0E22A50
Content-Type: text/html
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3DUS-ASCII">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
5.5.2653.12">
<TITLE>RE: [Veritas-bu] Incremental RMAN backups on split BCV's</TITLE>
</HEAD>
<BODY>

<P><FONT SIZE=3D2>The bottleneck is at the tape itself.&nbsp; We still =
have DLT7000-based libraries and doing nightly filesystem backups of =
what may become 3TB of data will take too long.&nbsp; We're going to =
have to move to RMAN backups so we can start doing incremental backups =
of the database files.&nbsp; So the problem is not the data source but =
the destination.</FONT></P>

<P><FONT SIZE=3D2>The attraction of doing RMAN backups from BCV's is =
two-fold.&nbsp; One, I'm off-host - there will be no load added to the =
production database servers from the backup process.&nbsp; Two, the BCV =
itself makes a good backup.&nbsp; Should I lose the DB while the BCV's =
are split, I can &quot;reverse establish&quot; (restore) the BCV's over =
the primary copies, roll it forward, and be back up and running without =
spinning tape.&nbsp; </FONT></P>

<P><FONT SIZE=3D2>A full RMAN recovery of the DB from tape can be done =
by fully restoring the tapes to the BCV's mounted on the media server, =
then synchronizing the BCV's against the primary mirrors.&nbsp; Still =
uncertain is if I can restore an RMAN backup on my media server =
directly to the primary database server.&nbsp; This latter move would =
be necessary for a single-tablespace restoral rather than a full DB =
restoral.</FONT></P>

<P><FONT SIZE=3D2>-M</FONT>
</P>

<P><FONT SIZE=3D2>-----Original Message-----</FONT>
<BR><FONT SIZE=3D2>From: Darren Dunham [<A =
HREF=3D"mailto:ddunham AT taos DOT com">mailto:ddunham AT taos DOT 
com</A>]</FONT>
<BR><FONT SIZE=3D2>Sent: Tuesday, January 21, 2003 3:38 PM</FONT>
<BR><FONT SIZE=3D2>To: Veritas-bu AT mailman.eng.auburn DOT edu</FONT>
<BR><FONT SIZE=3D2>Subject: Re: [Veritas-bu] Incremental RMAN backups =
on split BCV's</FONT>
</P>
<BR>

<P><FONT SIZE=3D2>&gt; </FONT>
<BR><FONT SIZE=3D2>&gt; The problem is that anticipated growth will =
make this difficult in the</FONT>
<BR><FONT SIZE=3D2>&gt; future - we won't be able to move enough data =
in the necessary</FONT>
<BR><FONT SIZE=3D2>&gt; timespan.</FONT>
</P>

<P><FONT SIZE=3D2>Can I ask what the bottleneck will be?&nbsp; I would =
think that the</FONT>
<BR><FONT SIZE=3D2>BCV/filesystem backup would have higher throughput =
than an RMAN backup.</FONT>
</P>

<P><FONT SIZE=3D2>&gt; Therefore, we're looking at using Oracle RMAN to =
perform database</FONT>
<BR><FONT SIZE=3D2>&gt; incrementals.&nbsp; However, I'm still wanting =
to perform the backup off-host so</FONT>
<BR><FONT SIZE=3D2>&gt; the idea is to use RMAN to backup BCV's mounted =
on my media server.</FONT>
<BR><FONT SIZE=3D2>&gt; </FONT>
<BR><FONT SIZE=3D2>&gt; The idea is this, and I'm looking for =
proponents or detractors to this idea:</FONT>
<BR><FONT SIZE=3D2>&gt; </FONT>
<BR><FONT SIZE=3D2>&gt; 1. split &amp; import as above to our media =
server</FONT>
<BR><FONT SIZE=3D2>&gt; 2. Open the database on the media server</FONT>
<BR><FONT SIZE=3D2>&gt; 3. perform the RMAN full and/or incremental as =
necessary</FONT>
<BR><FONT SIZE=3D2>&gt; 4. shutdown the Oracle instance</FONT>
<BR><FONT SIZE=3D2>&gt; 5. incrementally re-establish.</FONT>
<BR><FONT SIZE=3D2>&gt; 6. rinse &amp; repeat the next night</FONT>
<BR><FONT SIZE=3D2>&gt; </FONT>
<BR><FONT SIZE=3D2>&gt; The thought is that since the instance on the =
media server is opened &amp;</FONT>
<BR><FONT SIZE=3D2>&gt; closed normally and since the BCV should be a =
blockwise identical copy of</FONT>
<BR><FONT SIZE=3D2>&gt; the original database, then RMAN doesn't know =
from run-to-run that this is a</FONT>
<BR><FONT SIZE=3D2>&gt; &quot;new&quot; copy of the DB.&nbsp; As far as =
RMAN is concerned, the only DB involved</FONT>
<BR><FONT SIZE=3D2>&gt; lives exclusively on the media server - it =
doesn't have to know about the</FONT>
<BR><FONT SIZE=3D2>&gt; original mounted on the main server.</FONT>
</P>

<P><FONT SIZE=3D2>Sounds workable to me, but I'd want to test the =
restore.&nbsp; Do you have to</FONT>
<BR><FONT SIZE=3D2>have the other machine involved, or could you =
restore directly to the</FONT>
<BR><FONT SIZE=3D2>main server?</FONT>
</P>

<P><FONT SIZE=3D2>-- </FONT>
<BR><FONT SIZE=3D2>Darren =
Dunham&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ddunham AT taos DOT com</FONT>
<BR><FONT SIZE=3D2>Unix System =
Administrator&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Taos - The =
SysAdmin Company</FONT>
<BR><FONT SIZE=3D2>Got some Dr =
Pepper?&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;&nbsp;&nbsp;&nbsp; San Francisco, CA bay area</FONT>
<BR><FONT SIZE=3D2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =
&lt; This line left intentionally blank to confuse you. &gt;</FONT>
<BR><FONT =
SIZE=3D2>_______________________________________________</FONT>
<BR><FONT SIZE=3D2>Veritas-bu maillist&nbsp; -&nbsp; =
Veritas-bu AT mailman.eng.auburn DOT edu</FONT>
<BR><FONT SIZE=3D2><A =
HREF=3D"http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu"; =
TARGET=3D"_blank">http://mailman.eng.auburn.edu/mailman/listinfo/veritas=
-bu</A></FONT>
</P>

</BODY>
</HTML>
------_=_NextPart_001_01C2C1A6.D0E22A50--

<Prev in Thread] Current Thread [Next in Thread>