Re: PostgreSQL win32 fragmentation issue

On Sat, 2006-12-02 at 11:05 -0600, Andrew Dunstan wrote:
> Joshua D. Drake wrote:
> >
> > O.k., doesn't this seem like killing a squirrel with a 50mm tank gun?
> > The fact that this was suggested as anything but a last resort isn't to
> > inspiring.
> >
>
> Before you start ringing alarm bells, you need to instrument the actual
> performance effect.

Alarm bells? I saw a potential issue, I brought it to the community.
That is all.

> So far all I have seen is an assumption that the
> effect will be serious. Let's see some performance metrics that
> demonstrate the problem you think might exist.

It is not an assumption that defragmentation causes performance issues.
The assumption is that the above fact will cause problems with
PostgreSQL performance.

Yes, the PostgreSQL portion is an assumption which is why I have brought
it to the community for further analysis.

>
> The Windows port has been out for nearly 2 years. The fact that we have
> not seen complaints about this leads me to be somewhat skeptical.

Fair enough but I don't hold too much weight to this argument as even
the remote idea that anything but the slimmest <1% of our Wndows users
even read this list (let alone post to it) beyond a possible Google
search is far fetched.

> You
> could be right, but I want hard evidence.

Sure, which is why I opened the dialog.

> (And why wouldn't this be a
> problem for any DBMS running on Windows? There are just huge numbers of
> 24/7 Windows servers running SQLServer or Oracle.)

Well I don't know about SQLServer but Oracle uses a single file (at
least it does on Unix). I don't know if that would make a difference or
not.

Sincerely,

Joshua D. Drake

>
> cheers
>
> andrew
>
>
--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate