Re: Vacuum Full

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: Kevin Grittner <Kevin(dot)Grittner(at)wicourts(dot)gov>
Cc: Tino Schwarze <postgresql(at)tisc(dot)de>, pgsql-admin(at)postgresql(dot)org
Subject: Re: Vacuum Full
Date: 2009-04-02 14:14:52
Message-ID: 20090402141452.GF3163@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Kevin Grittner wrote:

> Yeah, I can't remember ever being glad I tried a VACUUM FULL. I think
> I once allowed it a week and a half to try to finish in a situation
> where there wasn't enough disk space for a rewrite (CLUSTER or the
> clever side-effect trick with ALTER TABLE). I even dropped all
> indexes and the primary key first, since I figured I would just need
> to rebuild them again if they were there, and it would slow down the
> VACUUM FULL. In the end, we killed it and used pg_dump and psql.
> That ran in about a day.

Yah. The worst part about VACUUM FULL is that if you interrupt it in
the middle, it bloats the table even more and nothing is cleaned up at
all. Worse, if you interrupt it at the wrong time (after the first
phase is finished I think), you can even get your server to PANIC, which
is not something that we allow any other command to do AFAIK.

--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Kenneth Marshall 2009-04-02 14:50:25 Re: postgresql 8.4
Previous Message Kevin Grittner 2009-04-02 14:06:10 Re: Vacuum Full