Re: Using Advance Replication for updating content from one server to the other...

For option 3, consider Materialized Views. Create them in your production system as a select on the staging tables, with whatever where clause takes your fancy. DML in staging doesn't so much get pushed to production as pulled across whenever the MV is refreshed. Nice.

(PS... don't cross-post. It's not appropriate, and annoys a lot of people.)

Regards
HJR

--
----------------------------------------------
Resources for Oracle: http://www.hjrdba.com
===============================


"Christian Sylvestre" <csyl_at_okami.com> wrote in message
news:83545ff7.0202150549.1e92f5ca_at_posting.google.com...


> Hello. I am looking for a good solid solution for a little/big problem

> ;-)

>

> I have 2 Oracle DB server running on 2 diffrent machines.

>

> The first DB server is used as the "staging/publishing" server. I.e.

> content updates are made on that server and approved on that server.

>

> We then need to push these content updates to the production DB

> server.

>

> However this need to be:

> - Done on demand.

> - Only a subset of the data of the "staging/publishing" server needs

> to be pushed to the production DB server.

>

> Here are the 3 solutions we looked at:

>

> 1- Trapping every DML (Update/Delete/Insert) done on the application

> level and writing them to a "kind" of script. Then copying that script

> from the "staging/publishing" server to the production DB server and

> then running that script with sqlplus from the command line.

>

> 2- Using a Database comparison tool to compare the 2 DB and then build

> a script that could be applied to the production server.

>

> 3- Using Oracle Replication to replicate the subset of the data from

> the "staging/publishing" server to the production DB server.

>

> FYI: The two severs are running 8.1.7 on AIX and the production

> servers are using Oracle HA (High Availability) Parallele servers.

>

> Option 1 is feasible but has a lot of failure points. Opction 2 could

> work but we have not found any command line comparison tools for

> UNIX...

>

> We are in favour of option 3 but we are not sure if Advance

> Replication would allow us to do so. I know that you can actually

> replicate only specific tables, but is it possible to replicate

> specific data in a table (based on a defined query)?

>

> Any comments/idea (even crazy ones) would be welcome!!!

>

> Cheers,

>

> Christian

>