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-devel

[Xen-devel] httpd persormance in Xen environment.

To: Xen Devel Mailing List <xen-devel@xxxxxxxxxxxxxxxxxxxxx>
Subject: [Xen-devel] httpd persormance in Xen environment.
From: nayden <nayden@xxxxxxxxx>
Date: Fri, 3 Dec 2004 06:48:35 -0700
Delivery-date: Fri, 03 Dec 2004 13:49:55 +0000
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=CAf0hzQ7F+oAy7SD5dBhZjzSKdqgdV29RV8kdYdZBKPVeW7YKMY56mHCYhyFBUatov0TzsJLEb/80H1DVPPm0QI6EfqwdufwNO44SVL8+fW7lgez1Xwc4OLYmiMpd7ForFkY+VK9aXZrjNgkhrjd6jPD/l/4wTFXWBTipmlWHnM=
Envelope-to: xen+James.Bulpin@xxxxxxxxxxxx
List-archive: <http://sourceforge.net/mailarchive/forum.php?forum=xen-devel>
List-help: <mailto:xen-devel-request@lists.sourceforge.net?subject=help>
List-id: List for Xen developers <xen-devel.lists.sourceforge.net>
List-post: <mailto:xen-devel@lists.sourceforge.net>
List-subscribe: <https://lists.sourceforge.net/lists/listinfo/xen-devel>, <mailto:xen-devel-request@lists.sourceforge.net?subject=subscribe>
List-unsubscribe: <https://lists.sourceforge.net/lists/listinfo/xen-devel>, <mailto:xen-devel-request@lists.sourceforge.net?subject=unsubscribe>
Reply-to: nayden <nayden@xxxxxxxxx>
Sender: xen-devel-admin@xxxxxxxxxxxxxxxxxxxxx
Hi, 

I am stresstesting httpd (Apache 1.3.31) under native linux and
xenolinux on order to evaluate the server performance in both
environments.  The hardware for each subsequent runs is identical, and
I also tried to keep the software configuration as similar as possible
(same versions of packages installed, same services enabled/disabled
etc.)  To test the performance of httpd I am using httperf with either
persistend and non-persistent connections.  When persistent
connections are enebled, I get very close performance on xenolinux to
native Linux, but when they are disabled the server performance over
xenolinux really suffers.  To give you a rough estimates how bad it is
here are some numbers.

persistent enabled   reply rate native                  xenolinux
4000 req/sec                   3900 rep/sec                700 rep/sec
persistent disabled         
4000 req/sec                   3950 rep/sec                3950 rep/sec

I understand why the case of persistent connections performs better
than the non-persistent, however in the former case httpd performance
under xenolinux is buffeling to me.

I suspect it is due to misconfiguration rather than xen issue and I
will appreciate if some of you provide any advices on what needs to be
checked.

The modifications that I have made to the kernel configuration of
linux, dom0 and dom0 are the following
# enable tcp time-wait recycling
echo 1 > /proc/sys/net/ipv4/tcp_tw_recycle

#disable tcp timestamps 
echo 0 > /proc/sys/net/ipv4/tcp_timestamps

# increase the number of local ports that are available for use
echo 5000 65535 > /proc/sys/net/ipv4/ip_local_port_range

max number of open file descriptors for each processor is increased to 32K. 

Native linux runs FC2 2.8.1 kernel that was compiled with different
configuration than  2.6.8.1-xen0 and 2.6.8.1-xenU. (perhaps the first
place to check?)

nayden


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/xen-devel

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