Re: Databse File layout on only 4 drives Ideas?

"Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message news:<h6pZ9.34797$jM5.89171_at_newsfeeds.bigpond.com>...
> "Joel Garry" <joel-garry_at_home.com> wrote in message
> news:91884734.0301271412.19f50e14_at_posting.google.com...
> > Mark Townsend <markbtownsend_at_attbi.com> wrote in message
> news:<BA5775D0.6EBA%markbtownsend_at_attbi.com>...
> > > in article AepY9.54719$L47.8070994_at_read2.cgocable.net, David Platt at
> > > david-platt_at_cogeco.ca wrote on 1/24/03 9:26 PM:
> > >
> > > > I am quite curious as to why a couple of you have written off the idea
> of
> > > > splitting data and index across drives. This is a practise that I
> have
> > > > followed for a while and I am wondering why it is being written off so
> > > > quickly
> > >
> > >
> > > First check the following -
> > > http://otn.oracle.com/deploy/availability/pdf/oow2000_same.pdf
> >
> > Which says: "Any access hot-spot that is smaller than a megabyte
> > should fit comfortably in the database buffer cache. Therefore it
> > will not create a hot-spot on disk."
> >
> > I've always had a problem with this statement. Wouldn't a db buffer
> > cache hot-spot like this be a "dirty write" attractor, often flushing
> > and making a hot-spot on the disk?
>
> No, because there's no relationship between the adjacent-ness of blocks on
> disk and what DBWR decides to flush.

Even in the case of, say, doing an update on everything in a partition that takes an hour during an otherwise low-usage time?
>
> (Which, I hate to mention it, explains why indexes and tables don't
> necessarily have to be separated. They might be read together (or close in
> time together, anyway), but that doesn't mean they will be *written*
> similarly).
>
> I didn't read the original link, however, so I may have taken your comments
> out of context.

I think many of us would be interested in your comments, especially about the part on disk file placement.

My own feeling on SAME is that it is inappropriate for large environments, but probably perfectly fine for places that won't spring for a DBA. I think the papers biggest mistake is in thinking everything is fine just because the methodology hides disk problems in cpu binding.

I agree that many DBA's spend too much tuning time focusing on useless "tuning." I think the SAME methodology goes too far the other way.

jg

--
@home is bogus.
"I LOVE that cowbell!"  Christopher Walken, in Blue Oyster Cult skit
on SNL.