Re: block chg tracking

On 02/12/2015 10:45 AM, MARK BRINSMEAD wrote:
> I don't see the connection between 24x7x365 uptime and level-0 backups.

Recovery times, specified in the service level agreements are much, much shorter with such stringent requirements. At least, that has been my experience.

>
> Yeah -- you're going to use a physical standby as your first line of
> recovery. Nobody is going to take their retail site offline for 4
> hours while you spin backup tapes. But in the event of a disaster,
> you are probably going to failover (immediately) to the standby and
> continue operations while you restore the (former) "primary" database
> from backups.
>
> Why would this preclude the "traditional" level-0 / level-1 backup?
>

Because it takes too long. While you are rebuilding the primary, you are exposed and have a single point of failure. So it must be done fast, probably through the snapshot revert technology. And with 50 TB database, 4 hours is a highly optimistic prediction, which requires some fancy hardware in place. With the traditional full/incremental cycle, the recovery would likely go on for a week. And SLA usually specifies hours, not weeks.

-- 
Mladen Gogala
Oracle DBA
http://mgogala.freehostia.com

--
http://www.freelists.org/webpage/oracle-l