Re: Assignment of undo tablespace to new user

DEspite the marketing, RAC doesn't scale awfully well. I'd say it's nearly linear one node to two nodes. After that, it tails off rather rapidly.

Sad, but true.

I'd still advocate it for high-availability. I'm not sure I'd promote it as a scalability/speed-up solution.

And I think that's intrinsic to inter-node communication issues. I can't see how you'd go about improving what's already in RAC, and it strikes me as one of the most 'adult' solutions on the market (compared with OPS or, heaven forbid, Microsoft's clustering technologies, RAC is streets ahead of the competition). But it still won't scale.

Regards
HJR "Paul Brewer" <paul_at_paul.brewers.org.uk> wrote in message news:3ecd49b3$1_3_at_mk-nntp-1.news.uk.worldonline.com...
> "Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message
> news:lx9za.40128$1s1.566625_at_newsfeeds.bigpond.com...
> snip
> > Never mind that in a RAC database, each instance gets its own undo
> > tablespace. So a 32-node RAC will have 32 perfectly legitimately active
> undo
> > tablespaces.
> Howard,
>
> How is your 32 node RAC scaling? I'd be interested in real world
> experiences.
> I must confess I'm somewhat cautious about RAC, since attending Jonathan's
> last presentation.
>
> Regards,
> Paul
>
>
>
Received on Mon May 26 2003 - 05:33:59 CDT