RE: Running Production Oracle Databases on OVM

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Thu, 21 Mar 2013 17:01:27 -0400
Message-ID: <029501ce2677$421d6730$c6583590$_at_rsiz.com>

Most of my clients run V stuff for the purpose of running more V machines than they have physical servers.

Sharing resources (V) is a recipe for variable response time, which is a major (perhaps THE major) cause of complaints regarding production servers.

Ergo, most of my clients do not run V for production, because they don't share physical resources for production.

If you want a measurement for the OVM overhead on your machinery, then do the same test with and without OVM.

One thing I know you don't want to do: mix para-virtualized and fully virtualized machines on the same box.

mwf

-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of fmh
Sent: Thursday, March 21, 2013 2:39 PM
To: oracle-l_at_freelists.org
Subject: Re: Running Production Oracle Databases on OVM

any ideas? hard to believe no one has worked with this technology.

On Wed, Mar 20, 2013 at 10:02 AM, fmh <fmhabash_at_gmail.com> wrote:

> For those who are currently running production databases using OVM.
> Can you please share your experience. Specifically ...
> 1) Are you using Oracle hardware (ODA, Exadata)? If not, why?
> 2) Oracle states there is a 5% penalty due to virtualized IO. How did
> you find your IO service time/throughput on OVM? Numbers will be
appricated.
> 3) What AWR workload profile data do you use in deciding if a db can
> be safely hosted on OVM e.g. sql executes, LIO, PIO? May db size?
>
> Thank you all.
>
>

-- 

----------------------------------------
Thank you.


--
http://www.freelists.org/webpage/oracle-l


--
http://www.freelists.org/webpage/oracle-l

Received on Thu Mar 21 2013 - 22:01:27 CET