Fwd: Dataguard setup at DR site
Date: Sat, 25 Apr 2009 19:32:01 -0700
I understand this aspect of corruption. But if I can afford to have that corruption (in the sense that I can afford to rebuild those objects), can I not enable force logging in data guard implementation and still have a working model with the caveat that objects with nologging would be corrupted on the standby and I should be prepared to rebuild them when I open the standby.
- Forwarded message ---------- From: Mark Strickland <strickland.mark_at_gmail.com> Date: Sat, Apr 25, 2009 at 5:49 PM Subject: Re: Dataguard setup at DR site To: howard.latham_at_gmail.com Cc: ksmadduri_at_gmail.com, Oracle-L_at_freelists.org
If you don't force logging in the primary and you have non-logged transactions in the primary, those transactions will not get to the standby (because they're not in the logs, of course) and you'll have physical (or is it logical?) corruption in the standby. However, you won't know it until you open the standby and try to query tables where transactions are missing. That's why forced logging is a prerequisite for setting up a physical standby.
Learned this one the hard way.
On Sat, Apr 25, 2009 at 2:48 PM, Howard Latham <howard.latham_at_gmail.com> wrote:
> DG evolved from 'log shipping' and need not exist now. It does provide extra features and it does TRY and manage
> everything for you but you can have a 'data guard' setup without dataguard. I was tld itit sorts out network failures etc for you and
> probaly the tools it gives you are worth it. I imagine that auditors might prefer a proper DG setup ?
> Coments anyone?