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


[Xen-users] IO performance depends on number of DomU's?

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] IO performance depends on number of DomU's?
From: Danilo Godec <danilo.godec@xxxxxxxxx>
Date: Tue, 28 Apr 2009 15:09:12 +0200
Delivery-date: Tue, 28 Apr 2009 06:09:56 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
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>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Thunderbird (X11/20081227)

I was monitoring IO performance of my Xen server using 'iostat' on Dom0
and 'bonnie++'. With no DomU's running, I get about 40 MB/s write and 70
MB7s read - which seems OK for a SW RAID5 array with 6 SATA drives...

Running only one DomU doesn't change much - running bonnie++ on either
Dom0 or DomU shows almost the same result (again monitoring 'iostat' on

After I start 16 DomU's, the IO performance goes down the drain - I only
get 5-6 MB/s running bonnie++ in Dom0 - or any of the DomU's. What's
interesting is that 'iostat' on Dom0 also shows only 5-6 MB/s of IO - so
it doesn't look like the cumulative IO of all DomU's would saturate the
actual IO capability of the hardware...

Any ideas? Things to try?

Is it possible to dedicate one of the CPUs to Dom0 exclusively?


Xen-users mailing list

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