Re: Parrallel query, server, option...

In article <32FB9580.48A_at_healthfare.com>, JJ Reynolds <jjr_at_healthfare.com> writes
>Are the parallel query and server options the same? Also, if my
>database is primarily written to, is there an advantage to upgrading to
>the parallel option?

The are different things. The parallel server option is for clusters or masssively parallel systems and allows multiple instances from different nodes to simultaneously access the same database.

Parallel query makes use of multiple processors (either an SMP system or a parallel server system) to speed up certain operations. It is essentially a MIS/data warehouse enhancement. Index builds, large data loads with sql*loader can be improved, but it is mainly a query improver as its name implies.

>
>I read somewhere that to use the parallel query option, your datafiles
>had to be on raw devices... is this true? If so, why? I already have a
>running database on a 2 CPU system, and, if feasible, would like to
>upgrade it to use the parallel features of Oracle.
>

You certainly need to be on raw devices to use parallel server, but I'm fairly sure that is not true of parallel query. The reason is that Unix filesystem cannot support the kind of distributed locking needed for parallel server.

As far as I am aware, parallel query is now bundled with the RDBMS and there is no extra cost.

Your physical database design may need to change to get any significant benefit.

-- 
Jim Smith