Re: Slow Vacuum was: vacuum output question

From: "Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com>
To: "Dan Armbrust" <daniel(dot)armbrust(dot)list(at)gmail(dot)com>
Cc: "Alvaro Herrera" <alvherre(at)commandprompt(dot)com>, "Andrew Sullivan" <ajs(at)crankycanuck(dot)ca>, pgsql-general(at)postgresql(dot)org
Subject: Re: Slow Vacuum was: vacuum output question
Date: 2009-01-06 21:33:30
Message-ID: dcc563d10901061333o25ef8507ib6b625d70f3d54fc@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Tue, Jan 6, 2009 at 2:07 PM, Dan Armbrust
<daniel(dot)armbrust(dot)list(at)gmail(dot)com> wrote:
>
> Actually, the customer reported problem is that when they enable
> autovacuum, the performance basically tanks because vacuum runs so
> slow they can't bear to have it run frequently.

Actually this is kinda backwards. What's happening is that the vacuum
uses up so much IO that nothing else can get through. The answer is
to make it run slower, by use of autovacuum_vacuum_cost_delay, and
setting it to 10 or 20 and seeing if they can then run autovacuum
during the day without these issues.

Note that vacuum was improved a fair bit from 8.1 to 8.2 and even
moreso from 8.2 to 8.3.

> Though, perhaps they had bloated indexes before they started
> autovacuum, and it never fixed them. Perhaps it will behave properly
> if we do a reindex, and then enable autovacuum.

Definitely look at the cost_delay setting. Makes a huge difference.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Tom Lane 2009-01-06 21:36:14 Re: Slow Vacuum was: vacuum output question
Previous Message Francisco Figueiredo Jr. 2009-01-06 21:31:37 Re: Generic provider for .Net