RE: Hardware performance Was: Does anyone run test win vs. linux and can share how muchfaster is linux than windows, please?

Regardless of your OS, your I/O subsystem is MUCH more important to your database performance than the OS is. This is also where you would see the tremendous slowness of a badly written query - when you're running it on a slow I/O system. If you run a bad query on super fast I/O, or on cached buffers, then it might appear to perform just fine, but again, this is dependend on the I/O speed - not the OS. I think the main differences to consider between operating systems are stability, cost, manageability, in-house experience, flexibility, compatibility, etc. - not performance.  

Regards,
Brandon Allen    

-----Original Message-----

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]On Behalf Of John Kanagaraj Sent: Wednesday, August 10, 2005 12:05 PM To: 'jkstill_at_gmail.com'; VerreyB_at_telkom.co.za Cc: oracle-l_at_freelists.org; 'Stephane Faroult' Subject: Now: Hardware performance Was: Does anyone run test win vs. linux and can share how muchfaster is linux than windows, please?

Actually, this makes an interesting topic.... Our very own Stephane Faroult wrote an excellent article for SELECT Q2 '04 titled "How Hardware performance translates into Oracle Performance". In it he showed how a *badly* written query shows a tremendous increase in performance when the hardware is upgraded while a well written query shows only marginal improvement when running on better hardware... Now I am not invalidating Jared's tests, nor am I claiming that Linux is better than Windows, but just pointing out that there could be a different twist to this whole thing. Maybe Stephane can summarize the findings here?  

John Kanagaraj

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Jared Still Sent: Wednesday, August 10, 2005 7:22 AM To: VerreyB_at_telkom.co.za
Cc: oracle-l_at_freelists.org
Subject: Re: Does anyone run test win vs. linux and can share how muchfaster is linux than windows, please?

On 8/9/05, Billy Verreynne (JW) < VerreyB_at_telkom.co.za> wrote:

Anyone that thinks that deploying Oracle on Linux will result in a faster system than deploying Oracle on Windows on the same h/w platform, is sorely mistaken.

Disagree.

I've run an application database on two nearly identical servers, one running linux, one running windows.

These were not quite identical, as the windows box had more disk. Both boxes had the same type of physical disks. The linux box had 3 RAID1 volumes for everything.

The win32 box had dedicated RAID1 volumes for redo, RAID10 for data/index, separate RAID10 for archives.

Running the same data feeding the same application, the linux box was noticably faster. No, I didn't measure it, it was not important at the time.

The linux server was actually running 3 databases at the time.

As for the Perl script, run strace on it. It generates time() calls a million times.

--

Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist

Privileged/Confidential Information may be contained in this message or attachments hereto. Please advise immediately if you or your employer do not consent to Internet email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of this company shall be understood as neither given nor endorsed by it.

--

http://www.freelists.org/webpage/oracle-l Received on Wed Aug 10 2005 - 14:16:14 CDT