Re: RMAN v Control Files

On Mon, 17 Feb 2003 01:02:09 +0000, Bill & Mary Border wrote:

> If you're asking about the age-old question of whether
> or not to use an RMAN recovery catalog...
>
> Pros to using RCVCAT:
>
> - The concept of database "incarnations" only works if
> you are using an RCVCAT. We found out (the hard)
> way that this is a big deal. That is, if you use the
> controlfile instead of an RCVCAT, you won't be able
> to use "incarnations". A BIG DEAL
> - Some of the reporting is better
>
> Cons to using RCVCAT:
>
> - Increased support required (one or more additional
> instances to support).
> - If you have multiple DB's with the same DBID, each
> will require its own RCVCAT, which is not practical
> in our case.
>
> Write to me if you need more info.
>
> Bill
>

Bill:

RCVCAT was the Oracle 8.0 (eight zero) method of creating catalogs. It isn't current, and your information is out of date.

There *are* functional restrictions in 8i and 9i with not having a catalog database. The principle one is you can't do TSPITR (tablespace point in time recovery) unless you have a catalog database... but when was the last time you ever did a TSPITR??

As for the business of incarnations... I'm not sure what you're saying. That you can't do incomplete recoveries with RMAN without a catalog? Not true. Or that you can't restore to a prior incarnation after having issued a resetlogs unless you're using a catalog? Not a big deal, if so, because I can't really see why anyone would ever need or want to restore to a prior incarnation, given the data loss involved.

Your circumstances might be different, of course, in which case you qualify as a 'complicated backup needs' environment.

But in 9i, it really isn't an issue not to have a catalog, TSPIRs aside.

Regards
HJR Received on Mon Feb 17 2003 - 04:06:02 CST