hpuxrac wrote:
> On Jul 11, 5:41 pm, joel garry <joel-ga..._at_home.com> wrote:
>> On Jul 11, 12:46 pm, "rogergor..._at_gmail.com" <rogergor..._at_gmail.com>
>> wrote:
>>
>>
>>
>>> You should ideally be able to restore only a single datafile if that's
>>> what's missing using RMAN, it's much faster than a full restore if the
>>> DB is multi-GB in size.
>> So how do you restore just a single table that is a small one of many
>> in the datafile and hasn't had (or you want to get rid of the effects
>> of) transactions since the export? While RMAN may be able to recover
>> blocks and datafiles, in general the granularity is tablespace. So
>> how many tables do you have, and how many of them are in their own
>> tablespace? I tell ya, I do a lot of work where I'm just updating one
>> or a few tables out of thousands, and being able to get those to a
>> known state in a much larger database is one of my main uses for exp/
>> imp these days. It's a backup and restore with table granularity (or
>> smaller with the query option, used carefully). Not everyone is on
>> 10g yet, flashback is just another feature.
>>
>> The real important point is that a DBA know when to use what and why.
>> Depending on export for general production backup is undeniably bad.
>> Export and cold backups is obsolete, yet one can justify it at a
>> stretch (I think Alexander is in that situation). Daniel is saying
>> how to do it right, my disagreement is that it is more important to
>> come up with a proper SLA than to overgeneralize to the point where
>> you are saying RMAN is a backup to the exclusion of everything else.
>> The R in RMAN stands for recovery, and that is not the purpose of
>> every backup.
>>
>> jg
>> --
>> @home.com is bogus.
>> Sometimes it's good to be the "do-nothing."http://www.signonsandiego.com/uniontrib/20070711/news_1b11prgn.html
>
> Anyone who doesn't think about the viability of a periodic export for
> all their important databases as ( at least ) a complement to other
> parts of their database recovery strategy just isn't thinking
> correctly.
>
> Joel is dead on here as he often is.
No one is arguing against periodic exports. The argument is against
exporting only and pretending that doing so constitutes having a
backup.
--
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org
Received on Wed Jul 11 2007 - 20:25:25 CDT
