Re: [SQL] Questions about vacuum analyze

Boy this is fun! ;-}

I have upgraded to 6.5.2 as you suggest. Binaries are built and I am now
re-running my vacuum just to give my self a benchmark starting point.

Just wanted to let you know that, for perl5, when I tried to run "make
test" after having built the perl kit, I get the following:

cobra:/opt/pgsql/postgresql-6.5.2/src/interfaces/perl5 $make test
PERL_DL_NONLAZY=1 /usr/bin/perl -Iblib/arch -Iblib/lib
-I/usr/lib/perl5/i386-linux -I/usr/lib/perl5 test.pl
Pg::conndefaults ........ ok
Pg::connectdb ........... ok
$conn->exec ............. ok
$conn->errorMessage ..... ok
pqReadData() -- backend closed the channel unexpectedly.
This probably means the backend terminated abnormally
before or while processing the request.
make: *** [test_dynamic] Error 255

Am I better off re-writing my perl scripts in 'C'? I've been thinking
about it in hopes of boosting the performance.

--
Steven Wheeler
UNIX Engineering
(918) 292-4119

Tom Lane wrote:

> "Steven M. Wheeler" <swheeler(at)sabre(dot)com> writes:
> > DB: version 6.5
> > 33 hours [to vacuum] - WOW!
>
> Yipes. We did find and fix a glitch in memory allocation in vacuum
> (causing both slowness and excessive memory usage) between 6.5 and
> 6.5.2. You might try updating to see if it's gotten any better...
>
> regards, tom lane