From: | Greg Stark <gsstark(at)mit(dot)edu> |
---|---|
To: | Joshua Marsh <icub3d(at)gmail(dot)com> |
Cc: | pgsql-performance(at)postgresql(dot)org |
Subject: | Re: Hardware/OS recommendations for large databases ( |
Date: | 2005-11-18 05:17:15 |
Message-ID: | 87zmo2pmas.fsf@stark.xeocode.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
Joshua Marsh <icub3d(at)gmail(dot)com> writes:
> We all want our systems to be CPU bound, but it's not always possible.
> Remember, he is managing a 5 TB Databse. That's quite a bit different than a
> 100 GB or even 500 GB database.
Ok, a more productive point: it's not really the size of the database that
controls whether you're I/O bound or CPU bound. It's the available I/O
bandwidth versus your CPU speed.
If your I/O subsystem can feed data to your CPU as fast as it can consume it
then you'll be CPU bound no matter how much data you have in total. It's
harder to scale up I/O subsystems than CPUs, instead of just replacing a CPU
it tends to mean replacing the whole system to get a better motherboard with a
faster, better bus, as well as adding more controllers and more disks.
--
greg
From | Date | Subject | |
---|---|---|---|
Next Message | Luke Lonergan | 2005-11-18 06:07:54 | Re: Hardware/OS recommendations for large databases ( |
Previous Message | Greg Stark | 2005-11-18 04:58:45 | Re: Hardware/OS recommendations for large databases ( |