Re: Scalability - Mailing list pgsql-general

From Tom Lane
Subject Re: Scalability
Date
Msg-id 25022.984582881@sss.pgh.pa.us
Whole thread Raw
In response to Scalability  (Jan Ploski <jpljpl@gmx.de>)
List pgsql-general
Jan Ploski <jpljpl@gmx.de> writes:
> Do I assume correctly that CPU will be the earliest limiting factor
> for a database (if we have a reasonable amount of memory in the box)?

Depends on your query mix.  If you have a lot of updates then I'd
recommend spending plenty of money on fast disk drives and good disk
controller boards.  If it's mostly SELECTs then you don't need such fast
drives, as long as you have enough RAM to cache a good chunk of the
active database.  In that case the CPU probably would be the limiting
factor.

> Is anyone out there using such a load balancing scheme with PostgreSQL?

Hardware is so cheap these days that I'd think it makes more sense just
to throw more hardware at the problem... I think client-driven load
balancing will be a pain in the neck to maintain.  (There has been some
talk of actual server support for load balancing, but it's still a long
way off.)

            regards, tom lane

pgsql-general by date:

Previous
From: Fabien COUTANT
Date:
Subject: [Q] post-crash behaviour
Next
From: "pgsql-sql"
Date:
Subject: