Re: struggeling with replication

Frank,

Thanks again for bearing with me.

Through continued experimentation, I've ascertained that the CREATE MATERIAL VIEW statement seem to require sysdba authority, and tries to create a table ... thus the failure on my manual CREATE statement. I'd presume that if the db-link had been in place at the time of the import, the CREATE mv statements would have failed like my manual ones did, seeing as the import had already created the underlying table.

At this point I think I'm in pretty good shape. I've used OEM to extract the CREATE mv statements, and will pre-pend to each a corresponding DROP TABLE and DROP MATERERIALIZED VIEW statement. Those can then be run immediately after each import cycle. I've done this manually to prove it out, and everything seems to be ok with it. The connectivity seems to be ok as well. FWIW, the import is also bringing in a lot of tables that are NOT associated with an MV, so we are still reliant on the import to do the basic production cutover load.

I think the only open question I have at this point (maybe there are questions I *should* ask and don't know to!) are with the REFRESH ... START WITH clause. It looks as if the date value in my extracted DDL is as of the time of the ddl extract. Should this be manually adusted?  Come time for the production cutover, the master site is going to be exported and moved about 48 hours before the MV site. Received on Tue Apr 25 2006 - 13:21:48 CDT