Message-ID: <F001.004E8328.20021014094837@fatcity.com>
Date: Mon, 14 Oct 2002 09:48:37 -0800
To: Multiple recipients of list ORACLE-L <ORACLE-L@fatcity.com>
Sender: root@fatcity.com
Reply-To: ORACLE-L@fatcity.com
Errors-To: ML-ERRORS@fatcity.com
From: becker.bill@marshfieldclinic.org
Subject: DW setup
Organization: Fat City Network Services, San Diego, California
Precedence: bulk
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit


Hello,

BACKGROUND:
We've been planning a 300GB datawarehouse architecture for Oracle 9.2 on
Solaris, and have proposed the following:
1) 2 separate instances of Oracle 9.2,
   - Instance A will be the staging instance, all ETL processing will
     take place here
   - Instance B will be the query instance, all reporting activity will
     take place here
2) Once data has been transformed, copy the tablespace metadata from 
   Instance A to Instance B using transportable tablespaces feature
3) The physical datafiles will be fast-copied from Instance A to Instance B
   using a vendor feature called Checkpoints (not Oracle's definition of checkpoints).

Point 3 needs further explanation: both of these instances will be connected to
Network Attached Storage (NAS) from a vendor named Procom. They have a feature
called Checkpoints, which quickly creates a read-only copy of a data volume
(I believe this is similar to network appliance's snapshot feature, and EMC also
has something like this, but the name escapes me). Checkpoints are very fast to
create, and can result in a read-only copy of 200GB of data in 1 - 2 minutes.
At present, we use them for backup purposes only, and they work well.

Instance A, the staging instance, will use the read-write Oracle datafiles
located on the procom read-write volumes. Instance B, the query instance,
will use datafiles located on a read-only procom volume, which also happens to
be the checkpoint volume of the read-write volume used by Instance A. The
checkpoint volume will be refreshed daily, from the staging volume, when the
daily ETL stream has completed. The query instance datafiles will be dropped
and re-created daily via the procom checkpoint, and the tablespace metadata 
will be plugged in using transportable tablespaces.

We have verified that Oracle works OK using plugged-in read only tablespaces
located on a procom read-only checkpoint volume.

QUESTIONS:

 (too much to hope for)
1) Is anyone else out there using this type of configuration with procom?
   If so, how well does this work? Any comments, problems?

 (more realistic)
2) Is anyone else out there using a similar configuration with a comparable
   vendor feature like checkpoints? Any performance problems? Any comments, problems?

 (more desperate)
3) Is anyone running a large Oracle data warehouse using primarily read-only 
   tablespaces? Any comments, problems? How do you refresh them?

 (last resort)
4) Does anyone care to comment on the above configuration? good idea...bad idea?

Thanks

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: 
  INET: becker.bill@marshfieldclinic.org

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru@fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).