Re: maintenance memory vs autovac - Mailing list pgsql-hackers

From Magnus Hagander
Subject Re: maintenance memory vs autovac
Date
Msg-id 49367B53.2050206@hagander.net
Whole thread Raw
In response to Re: maintenance memory vs autovac  ("Guillaume Smet" <guillaume.smet@gmail.com>)
List pgsql-hackers
Guillaume Smet wrote:
> On Wed, Dec 3, 2008 at 10:49 AM, Magnus Hagander <magnus@hagander.net> wrote:
>>> The autovacuum workers change that and make it a default behaviour (as
>>> we can have 3*maintenance_work_mem by default).
>> It's still one per process, it's just that autovac uses more than one
>> process.
>
> I agree. What I implied is that by default you have 3 autovacuum
> workers so the behaviour has changed, even if it didn't change in a
> technical way.
>
>> It's probably worthwhile to add a note about the effects of
>> autovacuum around the documentation of maintenance_work_mem, though.
>
> +1
> A lot of people set maintenance_work_mem quite high because of the old
> behaviour.

How about something as simple as this?

//Magnus
*** doc/src/sgml/config.sgml
--- doc/src/sgml/config.sgml
***************
*** 881,886 **** SET ENABLE_SEQSCAN TO OFF;
--- 881,891 ----
          than <varname>work_mem</varname>.  Larger settings might improve
          performance for vacuuming and for restoring database dumps.
         </para>
+        <para>
+         Note that when autovacuum runs, up to
+         <xref linkend="guc-autovacuum-max-workers"> times this memory may be
+         allocated, so be careful not to set the default value too high.
+        </para>
        </listitem>
       </varlistentry>


pgsql-hackers by date:

Previous
From: "Pavan Deolasee"
Date:
Subject: snapshot leak and core dump with serializable transactions
Next
From: "Fujii Masao"
Date:
Subject: Re: Sync Rep: First Thoughts on Code