Michael, if these are the only issues and your loads, being non-recoverable,
don't you have a recoverability issue if something breaks between your
backups. Or, do you just recreate and reload the damaged table(S) in that
case.

Still, have you investigated using incremental (full or delta) to ease the
workload and time between the full backups. Given 1TB+ of data, what
percentage of it does change from week to week?

The problem with using db2look to extract the ddl and then recreate the db
and then using loads is that you would have to drop the db if it got damaged
and recreate from scratch as you could not deal with hitting tblspaces
containers defined over the targets that running the output of db2look would
use.
The other exposure you may end up with is that you may not be able to
recover from a disaster by going to another system.
Dropping the db would then drop everything obviously, such as packages and
so on.  Do you use routines (stored procs, udfs,...) and how would you
recover those as db2look will only identify the source but not the compiled
routines.
Using full db backups (monthly say) and either full or incremental deltasat
the tblsp. level, nightly or weekly,  would certainly reduce the time and
volume used.  The issue then becomes how fast do you need to restore as
using incrementals may take you longer than reloading.  I'm not too sure
that reloading and therefore recreating all the indexes would necessarily be
faster than restoring from a full backup and then incrementals as only the
changed pages are handled in that case.

I'm just thinking here as I do not have a clear picture of how you're
operating.
HTH,  Pierre.

----- Original Message -----
From: "Michael Patterson" <[EMAIL PROTECTED]>
To: "db2users" <[EMAIL PROTECTED]>
Sent: Monday, March 10, 2003 3:56 PM
Subject: [DB2EUG] Restore vs Reload


> Hi All,
> I realize this question has been posted before, but I
> have not seen much in the way of response, so I am
> reviving the question.  I have a large data warehouse
> with controlled batch updates.  Currently I am doing
> full weekly backups, however it is very expensive in
> terms of time and resources.  Some data warehousing
> collogues  have recommended using a series of db2looks
> in combination with table exports to backup up my
> data.  Basically they are saying in a disaster
> situation, it would be faster to reload a table, or
> the whole database than to restore from tape.  I also
> have restore concerns since we use archive logging,
> but non-recoverable loads.  As a DBA I would prefer to
> have a regular full database backup, but see no reason
> to backup hundreds of GB's of indexes when they can be
> recreated.  Remember we store all load file after
> loading.  Does anyone use a similar unconventional
> backup strategy that they are willing to share?  For
> those who feel that need more specifics to comment,
> here they are.
>
> UBD-EEE 7.2 FP7
> AIX 4.3.3
> Tape backups
> Over a TB of data.
>
> I'm hoping to hear some thought on both sides of this
> argument.
> Thanks,
> Mike Patterson
>
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - forms, calculators, tips, more
> http://taxes.yahoo.com/
> -
> :::  When replying to the list, please use 'Reply-All' and make sure
> :::  a copy goes to the list ([EMAIL PROTECTED]).
> ***  To unsubscribe, send 'unsubscribe' to [EMAIL PROTECTED]
> ***  For more information, check http://www.db2eug.uni.cc
>


-
:::  When replying to the list, please use 'Reply-All' and make sure
:::  a copy goes to the list ([EMAIL PROTECTED]).
***  To unsubscribe, send 'unsubscribe' to [EMAIL PROTECTED]
***  For more information, check http://www.db2eug.uni.cc

Reply via email to