RE: Terabytes of archived redo logs

You mean everyone doesn't have spare servers sitting around waiting for DBAs to make good use of them? I'm shocked!

Seriously. that's why I asked if it were possible - meaning financially and logistically feasible. If it's not, then it's not. However, if possible, having a Standby for each Prod DB is not a horrible idea - as long as it's in another location.

I'd assume you'd use RMan to create the Standbys and continue to back up the Prod DBs. Guess there could be a potential recovery problem if you removed archived redo logs before RMan backed them up. However, that would be but one of the many challenges you'd face implementing any solution to this challenging problem.

Jack C. Applewhite - Database Administrator Austin (Texas) Independent School District 512.414.9715 (wk) / 512.935.5929 (pager)

 Same-Day Stump Grinding! Senior Discounts!

• Mike's Tree Service

"Mercadante, Thomas F \(LABOR\)" <Thomas.Mercadante_at_labor.state.ny.us> 05/03/2007 09:22 AM

To
<JApplewhite_at_austinisd.org>

cc
<avramil_at_concentric.net>, <oracle-l_at_freelists.org>,
<oracle-l-bounce_at_freelists.org>

Subject
RE: Terabytes of archived redo logs

How does Rman fit into this picture? Are you saying he should run a Physical Stand-by system just to fix a backup issue? It’s not a horrible idea. But I’m wondering what the cost-benefit-ratio would be. Acquiring a completely new server and stocking it with enough disk to hold the database and the archivelog files?  

Doesn’t sound like a reasonable financial solution to me.  

From: JApplewhite_at_austinisd.org [mailto:JApplewhite_at_austinisd.org] Sent: Thursday, May 03, 2007 10:11 AM
To: Mercadante, Thomas F (LABOR)
Cc: avramil_at_concentric.net; oracle-l_at_freelists.org; oracle-l-bounce_at_freelists.org
Subject: RE: Terabytes of archived redo logs  

Here's a, perhaps, wild, thought. Could you establish Physical Standby databases on (an)other server(s)? Then you could let your Prod datbases automatically shovel the archived redo logs to them, periodically remove them from the Prod environment as you see they've been transferred to the Standbys. You could also gzip them on the Standby side to further save space. Gzip is such a CPU hog that I'd not want it running on the Prod server.

You'd also get disaster recovery databases in the process. Just a thought.

Jack C. Applewhite - Database Administrator Austin (Texas) Independent School District 512.414.9715 (wk) / 512.935.5929 (pager)

Same-Day Stump Grinding! Senior Discounts!

• Mike's Tree Service

--
http://www.freelists.org/webpage/oracle-l