Date: Mon, 17 May 1999 12:06:08 -0600
Message-Id: <199905171806.MAA27409@earlj.nesusa.com>
From: Jason Earl <jearl@box100.com>
To: eg@tzv.fal.de
CC: pgsql-interfaces@postgreSQL.org
In-reply-to: <99051709490504.00692@salo.tzv.fal.de> (message from Eildert
	Groeneveld on Mon, 17 May 1999 09:45:49 +0200)
Subject: Re: [INTERFACES] Vacuum takes for ever
References: <99051709490504.00692@salo.tzv.fal.de>

   Hello everyone,

   in a substantial multi location development project we are using PG
   to create a large information system.  After having loaded data in
   the order of one million records a vaccum took 50h of CPU time on a
   350MHz PII (RedHat 5.2, pg 6.4.2). There were no deletes or anyting
   like that.

I made a test database when I was deciding whether PostgreSQL was
going to be up to the task.  It just so happens that it had a million
records in it as well.  The first vacuum took a very very long time,
subsequent vacuums, however, took much less time.

   What does vaccum do that takes so long and is there a way to speed
   this up?

Apparently it does some "database stuff."  Yikes, someone else will
have to answer that question.  I just know how to get data in and out
of PostgreSQL.

Jason Earl