I think you didn`t understand me.
I`ll run the full recovery once.
From now and on, I`ll continue to do an incr. backup from my source DB, and
on my target DB I`ll run an incremental recovery every day. I WONT OPEN THE
DB.
So, when I`ll decide to open this DB, I`ll get the latest version of my
source DB.
Afterwards, if I`ll want to run the process again - I`ll start with the full
recovery....
"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message
news:3fdfa787$0$18386$afc38c87_at_news.optusnet.com.au...
>
> "Ofer Razon" <orazon_at_012.net.il> wrote in message
> news:bro2if$aer$1_at_news2.netvision.net.il...
> > Hi.
> >
> > Let`s leave for the moment the issue of the PIT and the views.
> > These features are given by the Storage solution I use (SVM of
StoreAge).
> >
> > What I must understand is wheather the idea of taking incr. backups
every
> > day from my source DB and then make an incr. recovery to the target DB
> will
> > work.
> > I didn`t understand yet wheather it will work or not.
>
> I don't understand what's so difficult to understand about the word 'No',
> which I've used in my replies twice already.
>
> It won't work, because there is no such thing as an incremental recovery.
> Recovery is a one-off process, and can't be repeated or added to without
> blowing away the entire database and starting from scratch each time.
>
> Would scenarios help? Suppose you want to see a table as it was at 5.00pm
on
> Wednesday. It would be possible to restore Monday's full backup, apply
> Tuesday's incremental backup, and then perform incomplete recovery,
applying
> some of the redo generated throughout Wednesday, and thus open the
database
> as it was at 5.00pm on Wednesday. With the database open, you could
extract
> data as it was on Wednesday at 5.00pm.
>
> But you then want to see the data as it was at 10.00am on Thursday? Not a
> chance. You would have to destroy the Wednesday-at-5 database, re-restore
> Monday's full backup, re-restore Tuesday's and Wednesday's incremental
> backups, and then recover the database to the way it was at 10.00am on
> Thursday. You could then open the database and see the data as it was at
> that time.
>
> But then you want to see the data as it was at 1.00pm on Friday. Again:
> destroy the Thursday-at-10 database, re-restore the Monday full backup,
> re-restore the Tuesday, Wednesday and Thursday incremental backups, and
then
> perform a recovery until 1.00pm on Friday. The database could then be
opened
> and see the data as it was at 1.00pm on Friday.
>
> Each time you want a new time to open the "viewing" database, you have to
> destroy its prior version, and re-create it from scratch using the backups
> of, and redo logs generated by, the "production" database.
>
> I can't put it any plainer than that. Once a database has been recovered,
> that's it. If you want to recover it to a different time, you have to blow
> it away, and start from scratch. You can't apply further redo (ie, perform
> further recovery) to a database that has already been recovered and opened
> with a resetlogs.
>
> > I mean that on my target DB I`ll recover once a full recovery, and
> > afterwards I`ll run an incr. recovery everyday so when I`ll want to open
> the
> > DB it`ll contain the most lately version and data of my source DB.
>
> Again, there is no such thing as an incremental recovery. So, to put it
> plainly: "No this won't work".
>
> HJR
>
>
Received on Wed Dec 17 2003 - 12:29:02 CST
