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] Extremely slow disk access -- is there a way to optimize

To: Aquatic Bigfoot <aquaticbigfoot@xxxxxxxxx>
Subject: Re: [Xen-users] Extremely slow disk access -- is there a way to optimize?
From: Pasi Kärkkäinen <pasik@xxxxxx>
Date: Wed, 14 Oct 2009 19:45:56 +0300
Cc: xen-users@xxxxxxxxxxxxxxxxxxx
Delivery-date: Wed, 14 Oct 2009 09:46:39 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <4f600bcd0910140907o57315442r6004218028c98521@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: <4f600bcd0910140907o57315442r6004218028c98521@xxxxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mutt/1.5.13 (2006-08-11)
On Wed, Oct 14, 2009 at 09:07:51AM -0700, Aquatic Bigfoot wrote:
>    I've installed Xen on a new server with Intel Xeon E5520 CPUs and 24GB of
>    RAM, the OS is CentOS 5.3. I've copied over some virtual machines from
>    another Xen machine and they all seem to run a lot slower. The virtual
>    machines are allocated 512 RAM and one virtual CPU each. Running anything
>    that accesses the disk such as yum update or compiling drives the load up
>    and the virtual machine becomes unresponsive. On the previous Xen machine,
>    with a single CPU and 2GB of RAM I didn't run into this issue. The virtual
>    machines are file based, not LVM, and were file based on the previous
>    machine. I'm banging my head against the wall trying to figure out why
>    there is such a huge performance issue moving the virtual machines to a
>    beefier server.

1) Was the earlier server using the same OS/Xen versions? ie. CentOS 5.3
2) Have you benchmarked the storage subsystem of the new server? Does it
work OK from dom0? How is the performance compared to the old server? 
3) Have you checked "xm top" while domUs are doing IO?
4) Do you have iowait in dom0 while domUs are doing IO?

-- Pasi

Xen-users mailing list

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