Re: Q: Is Oracle 24*365 possible?

Hmm. Freudian slip that if ever I've seen one!

I meant write-write is addressed in the next release, of course.

Pete

Pete Sharman wrote:

> Jonathan
>
> I think you mean cache fusion rather than cold fusion don't you?
>
> I only mentioned it briefly because the original poster wasn't on the right
> release for it anyway. And of course, the read-read scenario is addressed in
> the next release. ;)
>
> Pete
>
> Jonathan Lewis wrote:
>
> > Cold fusion doesn't help very much though,
> > it only caters for consistent read blocks
> > being produced at the 'owner' rather than
> > pinging current blocks and rollback blocks
> > to the requester.
> >
> > Rapid update applications which don't
> > partition properly will probably be seriously
> > hit by pinging.
> >
> > --
> >
> > Jonathan Lewis
> > Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk
> >
> > Pete Sharman wrote in message <37E906D1.88973C57_at_us.oracle.com>...
> > >Dave
> > >
> > >Until you get to Oracle8, your application needs to trap the errors that
> > occur
> > >when the instance fails, then switch to the other instance. Applications
> > >developed with the Oracle8 OCI libraries can transparently failover.
> > >
> > >Not sure what you mean by the "common cause failure rate", but unless your
> > >application is designed specifically with OPS in mind, I wouldn't switch to
> > it
> > >without a good deal of up-front redesign work. This changes of course with
> > >Cache Fusion in 8i, but you're not on that release.

Received on Wed Sep 22 1999 - 14:59:06 CDT