From: | Francisco J Reyes <fran(at)natserv(dot)net> |
---|---|
To: | Chris Johnson <chris(at)chaska(dot)com> |
Cc: | rusland(at)scn(dot)ru, pgsql General List <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: FreeBSD: SMP and PostgreSQL |
Date: | 2003-02-10 02:43:08 |
Message-ID: | 20030209213700.P40450-100000@zoraida.natserv.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Thu, 6 Feb 2003, Chris Johnson wrote:
> So to the extent that your performance is limited by kernel "bandwidth"
> moving to FreeBSD 5.x will help. Chances are you'll be disk IO limited
> long before you get kernel limited, though.
I have to agree with Chris on this one on that the disk subsystem will
likely be a bigger player than switching from 4.X to 5.X
In my case I have a series of daily processes which I keep track or their
time of execution.
I looked at time improvements from going from a single CPU machine to a
dual CPU machine running Raid 5 and then the same machine running Raid 10.
Moving from single CPU to Dual CPU with Raid gave me somewhere in the 20%
improvement. However, when we switched to Raid 0+1 on the same SMP box I
saw performance jump to almost 3 times faster.
Although not exactly an answer to your question, but a testament that
a disk subsystem can be a real differentiator..
Also It realy would be best to wait until the 5.X branch has matured. I
may move my backup machine to 5 to give it a test/trial, but none of my
production machines are moving to it regardless of whether it is going to
be faster or not.
From | Date | Subject | |
---|---|---|---|
Next Message | Francisco J Reyes | 2003-02-10 02:44:43 | Re: How many joins is too many joins.... |
Previous Message | elein | 2003-02-10 00:36:06 | Re: PostgreSQL Mailing Lists in Italian? |