Re: Rollback in Postgres - Mailing list pgsql-sql

From Lewis Cunningham
Subject Re: Rollback in Postgres
Date
Msg-id 711191.77568.qm@web35601.mail.mud.yahoo.com
Whole thread Raw
In response to Re: Rollback in Postgres  (Simon Riggs <simon@2ndquadrant.com>)
Responses Re: Rollback in Postgres
List pgsql-sql
In addition to allowing you to read old data, Flashback will allow you to rollback to a point in time, including
returninga single table to a specific state.  Flashback database is like PITR without the log files.  
 

It started in 9i and improved dramatically in 10g.  11g has made additional improvements.

http://download.oracle.com/docs/cd/B19306_01/backup.102/b14192/intro007.htm

http://download.oracle.com/docs/cd/B28359_01/backup.111/b28270/rcmflash.htm


Lewis R Cunningham

An Expert's Guide to Oracle Technology
http://blogs.ittoolbox.com/oracle/guide/

Postgres Forums
http://postgres.enterprisedb.com/forum.do




--- On Fri, 7/11/08, Simon Riggs <simon@2ndquadrant.com> wrote:

> From: Simon Riggs <simon@2ndquadrant.com>
> Subject: Re: [SQL] Rollback in Postgres
> To: "Scott Marlowe" <scott.marlowe@gmail.com>
> Cc: "samantha mahindrakar" <sam.mahindrakar@gmail.com>, pgsql-sql@postgresql.org
> Date: Friday, July 11, 2008, 2:58 PM
> On Fri, 2008-07-11 at 11:21 -0600, Scott Marlowe wrote:
> > rollback after commit
> 
> Are you sure? 
> 
> Personally I don't think its viable. If it really does
> that it will
> would also need to rollback all transactions whose changes
> depend upon
> the earlier transaction. It would also need to track
> transactions that
> read data changed by an earlier transaction and then makes
> changes to
> the database. It's got no way to track that without
> extensive and costly
> additional infrastructure, since after transaction commit
> row locking
> information can be cleaned up by read-only transactions
> accessing those
> changed data blocks.
> 
> Flashback query allows reading data as it was at a certain
> point in the
> past. We might one day provide that, but undoing individual
> transactions
> isn't ever going to be feasible, without unknowable
> risk.
> 
> Not jumping on you, just think their marketing is ahead of
> the reality.
> 
> -- 
>  Simon Riggs           www.2ndQuadrant.com
>  PostgreSQL Training, Services and Support
> 
> 
> -- 
> Sent via pgsql-sql mailing list (pgsql-sql@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-sql


pgsql-sql by date:

Previous
From: Steve Midgley
Date:
Subject: Re: PERSISTANT PREPARE (another point of view)
Next
From: "Pavel Stehule"
Date:
Subject: Re: function that returns a set of records and integer(both of them)‏