Re: Allow sorts to use more available memory

From: Bruce Momjian <bruce(at)momjian(dot)us>
To: Stephen Frost <sfrost(at)snowman(dot)net>
Cc: Robert Schnabel <schnabelr(at)missouri(dot)edu>, Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>, pgsql-performance <pgsql-performance(at)postgresql(dot)org>
Subject: Re: Allow sorts to use more available memory
Date: 2011-10-05 21:54:17
Message-ID: 201110052154.p95LsHs02311@momjian.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Stephen Frost wrote:
-- Start of PGP signed section.
> * Robert Schnabel (schnabelr(at)missouri(dot)edu) wrote:
> > And getting back to the to-do list entry and reading the related
> > posts, it appears that even if you could set work_mem that high it
> > would only use 2GB anyway. I guess that was the second part of my
> > question. Is that true?
>
> Errr, and to get back to the to-do (which I've been considering doing
> something about...), it's to allow the *actual* memory usage for things
> like sorts to use more than 2GB, but as others have pointed out, you can
> do that by putting pgsql_tmp on a memory filesystem and letting the
> sorts spill to the memory-based FS.

It would be nice if the tempfs would allow us to control total temp
memory usage, except it causes a failure rather than splilling to real
disk.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Raghavendra 2011-10-06 09:31:23 Re: : Performance Improvement Strategy
Previous Message Ben Ciceron 2011-10-05 21:53:15 pg9 replication over WAN ?