Re: Incremental Backup and Change Block Tracking Questions

  • From: "Jared Still" <jkstill@xxxxxxxxx>
  • To: sbootsma@xxxxxxxxxxxxxx
  • Date: Fri, 22 Feb 2008 09:43:41 -0800

On Fri, Feb 22, 2008 at 7:52 AM, Sam Bootsma <sbootsma@xxxxxxxxxxxxxx>
wrote:

>  1. Have you found Incremental backups and Change Block Tracking to be
> reliable; i.e, can you reliably backup and then restore and recover?
>
> Incremental level 1 backups with BCT are much faster than level 1's
without BCT.
Restore is not a problem.
Level 1 backups are a very good thing.

Let's say you took a level 0 yesterday.
Last night a batch job loaded a table in direct path mode, which
will not appear in the archive logs.
Today you do a level 1 incremental.
The blocks loaded in direct path mode will be backed up by the level 1,
as they are recorded in the BCT.

So in the event of recovery, the level 1 backup not only skips all the
archive logs from the end of the level 0 to the beginning of the level 1,
the table loaded with direct path will also be recovered, which would
not have happened with the level 0 + archive logs.

>
>
> 2. What about Oracle compression during an RMAN backup?  Is that reliable?
>
>
Haven't heard any problems of reliability.  You may want to test it however,
as
my tests have shown quite a CPU hit.


> 3. How much of a performance hit using the Change Block Tracking feature?
> What about Oracle compression?
>
>
>
< 1% according to Oracle.  I haven't measured the actual CPU usage of
compression, only
its effect on a running database.


> 4. Do you know if any of these features are additional cost?  Do you know
> if there is a view in Oracle that outlines what features come with EE and
> what features are additional cost?
>
> Can't answer that.

>
>
> 5. We also have the option to do compression at the tape drive level.  Is
> anybody aware of any reason that Oracle compression (during backup) would be
> better or worse than letting the tape drive do it?
>
>
>
Tape level compression is probably faster.  A problem there might differing
implementations
of the compression between different versions of tape firmware.  Might be a
problem if you
need to restore old tapes with new hardware.

6. I have read that restores of a cumulative backup a generally faster than
> restores of a differential backup.  What makes cumulative faster?  Is it
> simply that RMAN will just one incremental file to restore and recover from
> (plus archive logs)?  If there are six differential backups to recover from,
> RMAN should be able to restore and recover the six files without DBA
> intervention.  I don't see a significant time difference.  Perhaps a couple
> minutes extra because it has six more files to deal with?
>
>
Haven't tested it, don't know.

-- 
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist

Other related posts: