Quoting Denis L'Huillier <dlhuillier AT PERSHING DOT COM>:

> I have a question..  We are currently going over our Disaster Recovery
> plans and will be doing a DR test in 2 weeks.  The major issue is
> restore
> times.  As a solution would it be possible to do "Incremental Restores"
> on
> the DR servers.  For example.. on the unix servers set up a crontab to
> run
> once a day doing a point-in-time restore from the production server.
> This
> way when a Disaster occurs the restore time will be as long as an
> incremental.  If we run the cron job after all the backups take place
> but
> before migration all the data should still be in the disk storage
> pool.
> This way we could do this for many DR servers at once..

A point in time restore will attempt to restore every file present on
the production server at the specified time, even if some of the files
are already present on the DR server. I think the fromdate and fromnode
options will do what you have in mind.

An "incremental restore" will not delete files from the DR server
because the corresponding files have been deleted from the production
server.

If you are going to have DR systems available full-time, you might
be better off with some form of remote disk mirroring that keeps the
DR system disk contents synchronized with the production system disk
contents. This would almost eliminate the loss of recent updates in
the event of a disaster.