Re: [HACKERS] Postgres Speed or lack thereof

>
> > > Since the bigger blocks are built on top of the existing
> > > memory context model, it would not conflict with any enhanced
> > > usage we could make with it.
> > >
> > > I include a patch at the end - please comment.
> > >
> > >
> > > Jan
> >
> > Did anyone play around with it? I've had it installed now for
> > some days and it work's well so far.
> >
> > How close are we to v6.5 BETA? Should I apply it to CURRENT?
>
>
> Good question. I am done with temp tables, so Vadim makes the call on
> when to start beta.

Now I'm placing another LOCK request (waiting for Vadim).

The problem with constraints in ExecRelCheck() I've just
found not only affects COPY. It also occurs with
INSERT...SELECT or big UPDATE's. Constraints with many
affected tuples eat up backend memory quickly until out of
swap space.

This fix must go into CURRENT and REL6_4 before we release
v6.5 or v6.4.3 either. There are actually reports from users
not able to reload dumps made with v6.4.2.

Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#======================================== jwieck(at)debis(dot)com (Jan Wieck) #