Re: Experiences with clusters on frequently joined data

Agreed:
We have a base bug open with the development team. The cluster works FINE until it grows too large, we haven't experimented to see what 'too large' is since we're already there.

Basically adding to the cluster is fine, but deletes on the cluster can screw up the entire cluster ... ouch. We have also noticed some 'strange' sql behaviour with the cluster.

How big is your cluster expected to get???

Thor HW

F Lange <flange_at_pt.lu> wrote in article
<01bc8c91$cd53d0a0$70d79ac2_at_LangeFrancois>...
> HI,
> We have used cluster to have better performances.
> We loadd the cluster early in the morning from a dowloaded file.
>
> during the day many queries.
>
> There is only on trouble solve in 7.4 Yes 7.4 with a back port patch
 fro
> 8.0.2
>
> The blocks are not freed if you delete all the rows in all the tables
> associate to a cluster , the block is still used and not freed.
>
>
> Regards Francois.
>
>
> P.S. other message will follow.
> Eric A. Stephens <estephen_at_javaboy.appliedtheory.com> wrote in article
> <m3en982rg1.fsf_at_javaboy.appliedtheory.com>...
> > Would anyone care to comment on the pros/cons of using clusters
> > to gain performance on frequently joined data. We certainly
> > want to avoid excessive denormalization in order to gain
> > performance and we view clustering as a viable alternative
> > --
> > Eric A. Stephens estephen_at_AppliedTheory.com
> > Software Engineering Group http://www.AppliedTheory.com
> > AppliedTheory Communications, Inc.
> >
>
Received on Wed Jul 09 1997 - 00:00:00 CDT