Re: CLUSTER

From: Adam Kavan <akavan(at)cox(dot)net>
To: "Claudio Lapidus" <clapidus(at)hotmail(dot)com>, <pgsql-general(at)postgresql(dot)org>
Subject: Re: CLUSTER
Date: 2003-10-24 15:45:40
Message-ID: 5.2.1.1.0.20031024101042.00a7edd0@pop.central.cox.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

At 02:02 AM 10/24/03 -0300, Claudio Lapidus wrote:
>Hello Adam
>
>I had a similar problem, although my table is much smaller, it keeps around
>50K rows, I think if you are inserting 10 by second you have 10*3600*24*7,
>that's about 6 million records. Also I'm assuming here that you also delete
>constantly any record being more than one week old, is this correct or do
>you delete all rows once a week and start over with the table empty?

Every night at midnight I delete everything over 7 days old.

>Essentially what we do here is vacuum very frequently (every five minutes or
>so), being very careful about FSM settings and run a VACUUM FULL and a
>REINDEX once a week, that keeps the table performance within limits. We
>never run CLUSTER for that.

I'm running pg_autovac on my database, and 7.4 (updating to beta 5 today to
see if it improves things).

--- Adam Kavan
--- akavan(at)cox(dot)net

In response to

  • Re: CLUSTER at 2003-10-24 05:02:24 from Claudio Lapidus

Browse pgsql-general by date

  From Date Subject
Next Message Network Administrator 2003-10-24 15:46:04 Re: HTML generation with PL/PgSQL
Previous Message Nailah Ogeer 2003-10-24 15:41:56 Semaphores