pg.ddx.io pgsql-admin@postgresql.org mailing list archive
help / color / mirror / Atom feedhardware to get best performance
4+ messages / 3 participants
[nested] [flat]
* hardware to get best performance
@ 2004-11-23 14:34 Vincent Chen <vctw@yahoo.com>
0 siblings, 1 reply; 4+ messages in thread
From: Vincent Chen @ 2004-11-23 14:34 UTC (permalink / raw)
To: pgsql-admin
Hi, all
Does postgresql support thread? Or let me put it this way, will postgresql get
better performance on multiple processor host than single processor? What's the
best platform, freebsd? linux?
I have been looking around, but still don't have answers for the above
questions.
Thanks,
Vincent Chen
-----------------------------------------------------------------
Yahoo!©_¼¯³y«¬ºëÆF
³Ì·sªº³y«¬ºëÆFñ¦WÀÉ¡AÅý«H¥ó¿W¨ãÓ¤H¦â±m¡I
http://tw.avatar.yahoo.com/
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: hardware to get best performance
@ 2004-11-24 02:36 Bruno Wolff III <bruno@wolff.to>
parent: Vincent Chen <vctw@yahoo.com>
0 siblings, 0 replies; 4+ messages in thread
From: Bruno Wolff III @ 2004-11-24 02:36 UTC (permalink / raw)
To: Vincent Chen <vctw@yahoo.com>; +Cc: pgsql-admin
On Tue, Nov 23, 2004 at 22:34:23 +0800,
Vincent Chen <vctw@yahoo.com> wrote:
> Hi, all
>
> Does postgresql support thread? Or let me put it this way, will postgresql get
> better performance on multiple processor host than single processor? What's the
> best platform, freebsd? linux?
If you have multiple connections to the database server making simultaneous
queries, multiple processors will be used. This is the normal case for
real database use.
The best platform for you is probably the one you are most familiar with.
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: hardware to get best performance
@ 2004-11-24 05:30 Vincent Chen <vctw@yahoo.com>
0 siblings, 1 reply; 4+ messages in thread
From: Vincent Chen @ 2004-11-24 05:30 UTC (permalink / raw)
To: pgsql-admin
>> Hi, all
>>
>> Does postgresql support thread? Or let me put it this way, will postgresql
get
>> better performance on multiple processor host than single processor? What's
the
>> best platform, freebsd? linux?
>
>If you have multiple connections to the database server making simultaneous
>queries, multiple processors will be used. This is the normal case for
>real database use.
>
>The best platform for you is probably the one you are most familiar with.
>
About platform, linux has much more filesystem choice than freebsd. If I put
postgresql data file on new filesystem like xfs or reiserfs, is it possible
that overall performance will be better than old filesystem like ufs?
Thanks for your response,
Vincent Chen
-----------------------------------------------------------------
Yahoo!©_¼¯³y«¬ºëÆF
³Ì·sªº³y«¬ºëÆFñ¦WÀÉ¡AÅý«H¥ó¿W¨ãÓ¤H¦â±m¡I
http://tw.avatar.yahoo.com/
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: hardware to get best performance
@ 2004-11-25 06:34 Jim C. Nasby <decibel@decibel.org>
parent: Vincent Chen <vctw@yahoo.com>
0 siblings, 0 replies; 4+ messages in thread
From: Jim C. Nasby @ 2004-11-25 06:34 UTC (permalink / raw)
To: Vincent Chen <vctw@yahoo.com>; +Cc: pgsql-admin
ext3 is just ext2 with journalling, and from what I've heard is rather
slow. ext2 is fast because it's async. RFS is fast at losing your data.
:)
UFS is rock-solid and is still being developed. In the past couble years
it's been changed to support some monsterous file and filesystem sizes
and has had soft-updates added. Soft-updates order metadata and data
writes in such a way that the filesystem will always come up in a
consistent state. It does what journaling does without the performance
drawback of a fixed journal space on the drive.
Another advantage FreeBSD has is that the kernel will take process
priority into account when it's scheduling disk IO. This means you can
nice long-running processes and not starve other processes for disk IO.
Some people think having a half-dozen filesystems to choose from is an
advantage. Personally, I'd much rather stick with a filesystem that's
proven itself for the past 30 years.
On Wed, Nov 24, 2004 at 01:30:06PM +0800, Vincent Chen wrote:
> >> Hi, all
> >>
> >> Does postgresql support thread? Or let me put it this way, will postgresql
> get
> >> better performance on multiple processor host than single processor? What's
> the
> >> best platform, freebsd? linux?
> >
> >If you have multiple connections to the database server making simultaneous
> >queries, multiple processors will be used. This is the normal case for
> >real database use.
> >
> >The best platform for you is probably the one you are most familiar with.
> >
>
> About platform, linux has much more filesystem choice than freebsd. If I put
> postgresql data file on new filesystem like xfs or reiserfs, is it possible
> that overall performance will be better than old filesystem like ufs?
>
>
> Thanks for your response,
>
>
> Vincent Chen
>
>
> -----------------------------------------------------------------
> Yahoo!?_???y?????F
> ???s???y?????F???W???A???H???W?????H???m?I
> http://tw.avatar.yahoo.com/
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
>
--
Jim C. Nasby, Database Consultant decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828
Windows: "Where do you want to go today?"
Linux: "Where do you want to go tomorrow?"
FreeBSD: "Are you guys coming, or what?"
^ permalink raw reply [nested|flat] 4+ messages in thread
end of thread, other threads:[~2004-11-25 06:34 UTC | newest]
Thread overview: 4+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2004-11-23 14:34 hardware to get best performance Vincent Chen <vctw@yahoo.com>
2004-11-24 02:36 ` Bruno Wolff III <bruno@wolff.to>
2004-11-24 05:30 Re: hardware to get best performance Vincent Chen <vctw@yahoo.com>
2004-11-25 06:34 ` Jim C. Nasby <decibel@decibel.org>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox