Re: How to replicate a database

Dino

You give any sizes.

If you really want the whole database replicated and its not massive, the best method is to recover a full (cold) backup regularly. This is easiest if the test database is on a different box to live (which is a good idea anyway): in this case you can mirror filesystems and directories completely. (It *is* a good idea to make simple changes once recovered (e.g. use different passwords) to ensure live isn't accessed by mistake)

If both are on one machine, you can still do it by recovering files to a planned set of different directories and do an ALTER DATABASE RENAME FILE for each file (you can script this).

If you only need to maintain certain key tables, why not develop exports for just those tables.

You can, of course, build in transaction level replication by using triggers and database links - but it will require development effort and could have a performance impact on live. V8 also has table replication but I haven't used this so I don't know what's involved.

Len

On Wed, 29 Sep 1999 11:04:42 +0800, "Dino Hsu" <dino1_at_ms1.hinet.net> wrote:

>In a project, to implement change management, we need to have a test
>environment as well as a production environment. The test database has the
>same data structure as the production database, the production database
>should be replicated to the test database once a day. Using import & export
>to replicate the database is very time-consuming, is there a better way to
>do it?
>We are using Oracle Database Server 8.0.5 for Windows NT.
>Please reply also to my mail box, thanks in advance.
>
>
>
>
Received on Sun Oct 03 1999 - 13:44:44 CDT