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] LVM over FC-San poor performance

To: Mice <mice1982@xxxxxxxxx>
Subject: Re: [Xen-users] LVM over FC-San poor performance
From: Jan-Aage Frydenbø-Bruvoll <jan@xxxxxxxxxxxxx>
Date: Tue, 19 Apr 2011 09:35:30 +0100
Cc: xen-users@xxxxxxxxxxxxxxxxxxx
Delivery-date: Tue, 19 Apr 2011 01:36:57 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <BANLkTintqw2X=zPtUbkho3OnNhmjUxfenA@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: <BANLkTintqw2X=zPtUbkho3OnNhmjUxfenA@xxxxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx

On Tue, Apr 19, 2011 at 09:26, Mice <mice1982@xxxxxxxxx> wrote:
> Hi, all
> We are about to test the performance with FC-SAN
> Currently there are 15 Hypervisors connected to a 16T LUN with 4G fiber 
> channel.
> LVM is used to manage the VM's disks, VM's disk image is created as logic 
> volume over the LUN.
> All VMs are HVM.
> when dd the lv from dom0, the performance is acceptable, range from 150 ~ 
> 300MB/s
> but when dd the local disk in guestOS (centOS 5.5 domU), the peformance is 
> poor, about 40MB/s
> my question is, is this normal? if not, any clues for performance tuning?

For what it's worth, I am struggling with a structurally similar
problem, although the underlying infrastructure is different. I have a
Nexenta storage server with COMSTAR iscsi set up - the performance is
acceptable from the dom0s (the network infrastructure becomes the
limiting factor before the server starts struggling), but as soon as I
benchmark the (paravirtualised) VM disks, I get similar numbers to
what you outline in your posting.

I am quite keen to find the source and a solution to this - if there
is anything I can contribute with along the way, please let me know.

Best regards

Xen-users mailing list

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