Re: Out of Memory - 8.2.4 - Mailing list pgsql-general

From Marko Kreen
Subject Re: Out of Memory - 8.2.4
Date
Msg-id e51f66da0708300654r2308bdfeu8b0d879be3e0b673@mail.gmail.com
Whole thread Raw
In response to Re: Out of Memory - 8.2.4  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
On 8/30/07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> "Marko Kreen" <markokr@gmail.com> writes:
> > Note that it's much better to err on the smaller values.
>
> > Extra index pass is really no problem.
>
> I beg to differ ...

Well, if Postgres tries to cut down passes by using max memory
then admin is forced to cut down maint_mem for safety reasons...

> What this may actually suggest is that autovacuum needs its own value of
> "maintenance_work_mem", or that it should automatically divide the value
> by the authorized number of workers.

+1

Autovacuum having it's own value and cutting pieces from there
and giving to vacuums is probably the easiest path.

Then the per-backend maint_mem does not need to be large.

--
marko

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Out of Memory - 8.2.4
Next
From: Alvaro Herrera
Date:
Subject: Re: Out of Memory - 8.2.4