Re: Test cluster slower than single machine

[Daniel Blumenthal]

> Regarding joins: in the current versions of mysql
> cluster, joins are not handled very well, this is
> something we hope to fix.
>
> More over, there are very likely lots of smaller
> adoptions of your program that will make it run
> faster on cluster, that does not make any
> difference for any other dbms or mysql storage
> engine.

I'm gearing up to switch from a single server over to a clustered server, and I'm curious if there's a place that describes the kinds of changes you should make to optimize SQL queries for clusters. I've read the Davies / Fisk book (which was great), but I don't remember this being explicitly covered. (other than a reference to joins being inefficient because they might require accessing data on another machine).

Any thoughts?

Daniel