From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
---|---|
To: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
Cc: | Jeison Bedoya <jeisonb(at)audifarma(dot)com(dot)co>, PostgreSQL mailing lists <pgsql-performance(at)postgresql(dot)org> |
Subject: | Re: Process in state BIND, authentication, PARSE |
Date: | 2013-07-09 13:23:39 |
Message-ID: | CAHyXU0wcbdUdZbiQXxzMFO3+=PRRP4NiH=PvPy-m1i0c84dYgA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
On Mon, Jul 8, 2013 at 5:35 PM, Michael Paquier
<michael(dot)paquier(at)gmail(dot)com> wrote:
> On Tue, Jul 9, 2013 at 2:01 AM, Jeison Bedoya <jeisonb(at)audifarma(dot)com(dot)co> wrote:
>> max_connections = 900
>> work_mem = 1024MB
>> maintenance_work_mem = 1024MB
> Aren't work_mem and maintenance_work_mem too high? You need to keep in
> mind that those are per-operation settings, so for example if you have
> 100 clients performing queries, this could grow up to 100G. In your
> case you even have a maximum of 900 connections... Do you perform
> heavy sort operations with your application that could explain such an
> amount of memory needed?
it's not at all unreasonable for maintenance_work_mem on a 128gb box.
agree on work_mem though. If it was me, i'd set it to around 64mb and
then locally set it for particular queries that need a lot of memory.
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | Josh Berkus | 2013-07-09 22:14:29 | Re: Performance autovaccum |
Previous Message | Michael Paquier | 2013-07-08 22:35:14 | Re: Process in state BIND, authentication, PARSE |