[Pgpool-general] PGPool and parallel query

[Tatsuo Ishii]

> Tatsuo,
> 
> Thanks for the help. I´ve looked at postgresforest and became interested
> about its table partitioning among servers. Right now, i guess that my
> problems are related with some nested loops because we have some joins and
> we´ll try to optimize the main query.
> 
> I´ve not thinked yet about going to FISL but it´s an interesting opportunity
> to meet you. So, i´ll try to be there.

I think someone who are going to FISL is Josh Berkus, not me.
--
Tatsuo Ishii
SRA OSS, Inc. Japan

> Regards
> Adonias
> 
> On 4/3/07, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> >
> > > Adonias,
> > >
> > > > I´m developing a large database inside Brazilian Government and we are
> > > > using postgresql for about 2 years as our DBMS.
> > >
> > > Cool!  I'm going to be at FISL and TechDays Sao Paulo; maybe we could
> > meet up?
> > >
> > > > As the most complicated query is made of 3 unions, i´d like to run
> > them in
> > > > parallel. Looking at PGPool II i´ve found the way to do parallel
> > queries.
> > > > But, my question is:
> > >
> > > Did you look at PostgresForest as well?  Ishii-san, what's the status of
> > > PGPoolII parallelization vs. PostgresForest?
> >
> > PostgresForest can only be used with Java applications. If his
> > application is implemented in Java maybe worth to look at.
> >
> > > > To use and take advantage of the parallel query implementation of
> > PGPool i
> > > > have to setup one postgresql node for each parallel execution line
> > that i
> > > > wanna have? Is it possible to pgpool start parallel execution based on
> > my
> > > > number of processors into the node?
> > >
> > > I don't know any parallel query version of PostgreSQL which does not
> > require a
> > > separate db note for each parallel execution node, Greenplum included.
> >
> > Right. You can run multiple PostgreSQL DB cluster on a machine to take
> > advantage of pgpool parallel execution however.
> > --
> > Tatsuo Ishii
> > SRA OSS, Inc. Japan
> >