From: | Culley Harrelson <harrelson(at)gmail(dot)com> |
---|---|
To: | pgsql-performance(at)postgresql(dot)org |
Subject: | Need suggestions on kernel settings for dedicated FreeBSD/Postgresql machine |
Date: | 2009-08-07 21:24:21 |
Message-ID: | 529825a30908071424i480489b8hc7463231301f99d6@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
Hi Everyone,
I manage a freeBSD server that is dedicated to postgresql. The
machine has 4 gigs of ram and there is a single database powering a
web application that is hosted on a neighboring machine. The web
application is mostly reading the database but there are considerable
writes and I don't want to tune the machine exclusively for writes. I
realize more information would be needed to optimally tune the machine
but I am seeking advice on making some sane kernel settings for a
general purpose database on a dedicated system. Currently I have:
$ cat /etc/sysctl.conf
kern.ipc.shmmax=268435456
kern.ipc.shmall=65536
and
$ cat /boot/loader.conf
kern.ipc.semmni="256"
kern.ipc.semmns="512"
kern.ipc.semmnu="256"
In postgresql.conf I have:
max_connections = 180
shared_buffers = 28MB
I would like to increase this to 256 connections and make sure the
kernel settings are giving postgresql enough breathing room without.
I suspect my settings are conservative and since the machine is
dedicated to postgresql I would like to give it more resources if they
could be used. Any suggestions?
culley
From | Date | Subject | |
---|---|---|---|
Next Message | Greg Stark | 2009-08-08 00:28:20 | Re: PG-related ACM Article: "The Pathologies of Big Data" |
Previous Message | Scott Carey | 2009-08-07 21:09:23 | Re: SQL select query becomes slow when using limit (with no offset) |