This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-users] how many VCPUs per physical CPU?

To: Peter Peltonen <peter.peltonen@xxxxxxxxx>
Subject: Re: [Xen-users] how many VCPUs per physical CPU?
From: Peter Booth <peter_booth@xxxxxx>
Date: Mon, 06 Jul 2009 10:09:50 -0400
Cc: xen-users <xen-users@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Mon, 06 Jul 2009 07:11:24 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <69b6a1580907060548x1a9c6ce3j43f327977c3706f8@xxxxxxxxxxxxxx>
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
References: <69b6a1580907060548x1a9c6ce3j43f327977c3706f8@xxxxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
The "right answer" depends on what you want to optimize for. Your webserver has human users for whom a 500ms delay has a measurable impact. How important are those users compared to snappy email?
If a vm has to wait, which vm should it be?
What workload, normal and peak, do you anticipate for each vm?

My prejudice, as someone who works with website performance is to maximize user experience, but only you know your business context.

I think that it's best to start simple and approach CPU tuning with a "trust but verify" mindset. I'd probably only dedicate a CPU for dom0, and

Sent from my iPhone

On Jul 6, 2009, at 8:48 AM, Peter Peltonen <peter.peltonen@xxxxxxxxx> wrote:

Hi list,

I have a machine with  Intel Xeon E5420 quad core processor and I'm
now wondering how to divide the CPU between all Xen instances.

I am running one qmail based mail server and several LAMP web servers.
The server OS is CentOS Linux.

What I've learnt from the archives is that dom0 should have its own
dedicated core.

But what about the rest, the mail server needs most CPU time, so I
could dedicate a core for it too. Would id benefit from having access
to the "pool" of the remaining cores too?

And what about VCPUs. Would a dom0 or domU benefit having say 2 VCPUs
and running an SMP kernel while in reality it is running only on 1
physical core?


Xen-users mailing list

Xen-users mailing list

<Prev in Thread] Current Thread [Next in Thread>