WARNING - OLD ARCHIVES

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/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-users

Re: [Xen-users] slow read IO in domU

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-users] slow read IO in domU
From: <desire@xxxxxxxxx>
Date: Sun, 15 Jan 2006 06:59:36 +0800
Delivery-date: Sat, 14 Jan 2006 23:06:43 +0000
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:references; b=B3YhQMGpkHfDSlQj/7GUBTzGz7L8rBxddisqTUzjrUZq29ILfgkIKyPQaF4JcFNUQJt0a53e2c+RXqX9aGK+nS+f8KyVMJXlKlVqUHaoOJSeR2V2R4PsNJXlyQ1wFF/WYMSwbuibD1Ce/XR64VqGqIKu2f+G+FRl1IqEAigVf+s=
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <bf77ac7b0601132304t42a0f932g59c6c53c5a66cbb6@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/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
References: <bf77ac7b0601132304t42a0f932g59c6c53c5a66cbb6@xxxxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
Sorry for the duplicate mail... gmail had some problems...

On 1/14/06, desire@xxxxxxxxx < desire@xxxxxxxxx> wrote:
Hi,  I think I must've done something incorrectly, but I can't quite figure out what it is.

To compare disk I/O performance in dom0 and domU, I ran the following two dd lines and observed vmstat 1 results (for blocks read/written per second), dd's reported speed, etc.  both dom0 and domU are allocated 512MB of RAM, so reading the 1GB file should not benefit from caching.

dd bs=1000000 count=1000 if=/dev/zero of=junk
(sync, wait, etc)
dd bs=1000000 count=1000 of=/dev/zero if=junk

In dom0, I achieve 45MB/s write speed and 63MB/s read speed reasonably consistently.  I guess this should be reasonable speeds for a partition on LVM on linux software raid 0 on a pair of SATA drives near the front of the disk.  This speed was observed using dd output as well as dom0 vmstat.

In domU, I achieve also about 45MB/s write speed but only about 13MB/s read speed consistently, while writing to a "sda3" residing on a dedicated LVM volume on software raid 0 on the same part of disks.  This speed was observed in vmstat on both dom0 and domU.

This is on xen 2.0.6 with linux 2.6.9 dom0 and domU kernels.  Shouldn't dom0 and domU achieve similar read & write speeds? Thanks for any pointers as to where I've gone wrong...


I've now noticed that I get different results running the same tests on my domU's sda1 (ext3 on a dom0 loopback file residing on a reiserfs-formatted LVM volume on software raid device) and sda3 (reiserfs on a dom0 LVM volume on software raid device).  Although sda1 should intuitively be slower since it goes through an additional loopback file, I achieved closer read/write speeds to dom0 (~ 42/40 read/write MB/s), while sda3 is noticeably slower at reading (~ 13/45 read/write MB/s).

Is this difference in speed likely to be caused by loopback file on LVM vs. direct LVM, or perhaps by ext3 vs reiserfs, or something else?


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
<Prev in Thread] Current Thread [Next in Thread>