People are running larger InnoDB databases on poorer hardware. Note
that I wouldn't dream of it because I care about data integrity and
stability, but this discussion is purely about performance and I know
it is possible.
I am sure throwing hardware at it is not the solution. Just trying to
highlight what the root cause is. Raptor disks are not that bad, even
if there's just "one" disk with RAID1, especially for a SELECT-heavy
web app.
Scott's idea of upgrading to 9.x is a good one. But it's not been easy
in the past. There have been issues related to UTF-8, after the whole
RPM stuff on CentOS has been sorted out.
QUESTION:
If auto_vaccum is ON, and I'm running a manual vacuum, will they
coflict with each other or will basically one of them wait for the
other to finish?
On Mon, Apr 18, 2011 at 2:13 AM, Alan Hodgson <ahodgson@simkin.ca> wrote:
> On April 17, 2011, Phoenix <phoenix.kiula@gmail.com> wrote:
>
>> >> Surely this is not tenable for enterprise environments? I am on a
>
>> >> 64bit RedHat server with dual CPU Intel Woodcrest or whatever that was
>
>> >> called. Postgres is 8.2.9.
>
>> >>
>
> .. and you have essentially 1 disk drive. Your hardware is not sized for a
> database server.
>
>>> it's a RAID 1 setup. Two Raptor 10000rpm disks.