Re: Parallel Query Processing

On Thu, 2009-03-26 at 10:12 -0700, David Fetter wrote:
> On Thu, Mar 26, 2009 at 05:04:29PM +0000, Simon Riggs wrote:
> >
> > On Wed, 2009-03-25 at 15:52 -0600, Scott Marlowe wrote:
> > > On Wed, Mar 25, 2009 at 2:24 PM, aravind chandu <avin_friends(at)yahoo(dot)com> wrote:
> > > > Hello,
> > > >
> > > > I have a few questions related to the parallel query
> > > > processing.Can you guys tell me how to implement parallel query
> > > > processing in postgresql database.
> > >

> > > PostgreSQL provides no capability to parallelize one query into
> > > multiple processes.
> >
> > I have a tool that will allow you to manually parallelize a query
> > (and get the right answer). That is useful for certain larger
> > queries. I'll be publishing that a bit more visibly in next few
> > months.
>
> Is this snapshot cloning?

Basically, yes.

> On the subject of snapshot cloning, I can see, at least in broad brush
> strokes, how this would work for read queries, but I'm curious how (or
> if) it could work for writes like, for example, updates of many
> partitions at once. Could it?

Yes, think so, since most writes involve first reading the data.

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support