Re: Recovery Manager (RMAN) problem

From: Vladimir M. Zakharychev <>
Date: Mon, 2 Jun 2008 05:46:52 -0700 (PDT)
Message-ID: <>

On Jun 2, 1:25 pm, sati_82 <> wrote:
> Hello,
> I have the following problem. I need to implement a mechanism which
> will restore a database to its state in the past. The restore
> operation should start on demand. The machanism is supposed to be some
> kind of antidote for user errors (e.g. a user had changed data in DB
> and after that it turned out that all changes made were wrong and DB
> has to be restored to its state before the user changes).
> For implementastion such mechanism I plan to use RMAN. I'm doing test
> for the moment and I found out that it doesn't work as I thought. I do
> the following:
> 1. I backup (with RMAN) a TABLESPACE on which database stores user
> data
> 2. I make changes in a table (the table is stored on backuped
> 3. I restore the TABLESPACE from the file backuped at point 1.
> The problem is that changes made (2) after I backup a tablespace (1)
> exists in the table (but I think it shouldn't because I restore the
> tablespace)(3).
> Could you please tell me what is the reason for that?
> I appreciate any help.
> Best regards,
> krys_tek

First of all, if you are running 10g, you can use flashback to restore tables to some arbitrary time in the past (within undo_retention limits.) No need to do full-blown PITR. For example, you can

CREATE TABLE my_table_as_of_before_the_bad_data  AS SELECT * FROM my_table AS OF {timestamp|SCN}

    {WHERE ...} and this will create a copy of the data as it was at the specified SCN or within +/-3 seconds of the timestamp. You can the replace the bad data from this copy or simply swap the tables.

Secondly, how did you restore your tablespace (what were the exact commands you entered and the output from RMAN you received?) Are you sure you did TSPITR?


   Vladimir M. Zakharychev
   N-Networks, makers of Dynamic PSP(tm) Received on Mon Jun 02 2008 - 07:46:52 CDT

Original text of this message