Re: Is vacuum full lock like old's vacuum's lock? - Mailing list pgsql-general

From Andrew Sullivan
Subject Re: Is vacuum full lock like old's vacuum's lock?
Date
Msg-id 20020302132032.A28043@mail.libertyrms.com
Whole thread Raw
In response to Re: Is vacuum full lock like old's vacuum's lock?  (Francisco Reyes <lists@natserv.com>)
Responses Re: Is vacuum full lock like old's vacuum's lock?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
On Sat, Mar 02, 2002 at 01:23:59AM -0500, Francisco Reyes wrote:
> On 1 Mar 2002, Neil Conway wrote:
>
> > Why do you need 'vacuum full' rather than just 'vacuum'? Read the docs,
> > it's usually not necessary.
>
> Because every night I delete/reload a big chunk of my data and then once a
> week I delete/reload the entire dataset... about 7 million records.

This is one thing that is slightly confusing (to me) about the new
vacuum -- perhaps someone more familiar with the internals can
clarify?

I thought that, in the case Mr Reyes is talking about, Postgres would
again use the freed disk space.  It's just that the space would not
be available to other applications.  I thought what VACUUM FULL did
was just free the disk space _absolutely_.

If I'm right, does that also mean that performance is actually
(marginally) _better_ in these types of cases, because the system
doesn't need to request new disk blocks from the OS?

A

--
----
Andrew Sullivan                               87 Mowat Avenue
Liberty RMS                           Toronto, Ontario Canada
<andrew@libertyrms.info>                              M6K 3E3
                                         +1 416 646 3304 x110


pgsql-general by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Clues about tables fileformat
Next
From: Tom Lane
Date:
Subject: Re: Is vacuum full lock like old's vacuum's lock?