[Pgpool-general] Pgpool and pg_autovacuum
Tatsuo Ishii
t-ishii at sra.co.jp
Mon Jul 4 13:57:14 GMT 2005
> Ok,
> And what about transaction ID? May VACUUM change ID's in a different way
> between master and slave, thus causing pgpool entering degeneration mode?
Different XIDs among servers do not immdediately cause the
degeneration. However if your queries depend on XIDs, that might be a
problem. For example, if your query is something like this:
SELECT * from t1 WHERE xmin > 10000;
In this case servers might return different number of result rows and
that will cause the degeneration.
--
Tatsuo Ishii
> Thanks.
>
> > -----Original Message-----
> > From: Tatsuo Ishii [mailto:t-ishii at sra.co.jp]
> > Sent: Sábado, 02 de Julio de 2005 01:43 a.m.
> > To: Pablo G. Milano
> > Cc: pgpool-general at pgfoundry.org
> > Subject: Re: [Pgpool-general] Pgpool and pg_autovacuum
> >
> > > Is it safe to use the pg_autovacuum daemon with pgpool?.
> > I've read in
> > > postgress manual (sect 21.1.3) that vacuum can modify the
> > transaction ID of
> > > old transactions in order to prevent transaction ID
> > wraparound. Does it
> > > interfere with pgpool health check?.
> >
> > I think no. What pgpool health check is doing is just connecting to
> > postmaster. So there should be no interaction with VACUUM.
> > --
> > Tatsuo Ishii
More information about the Pgpool-general
mailing list