From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hui Kang Subject: dom0 kernel v2.6.32.41 network problem Date: Fri, 15 Jul 2011 17:54:11 +0000 Message-ID: Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============0386370709==" Return-path: Content-Language: en-US List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org --===============0386370709== Content-Language: en-US Content-Type: multipart/alternative; boundary="_000_CA45F7026312hkangcsstonybrookedu_" --_000_CA45F7026312hkangcsstonybrookedu_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi, I am running two hosts with xen-4.0.1 and dom0 kernel 2.6.32.41. The two ho= sts are connected to a Gigbit switch. Then I use iperf to measure the bandw= idth and got the results of 200Mbit/second, which is only 1/5 of the availa= ble bandwidth. Then I boot these two hosts using a vanilla kernel of 2.6.32.34. The iperf = reports 937Mbis/s, which is very close to the available link bandwidth. So why the pvops dom0 kernel performs so bad? Does anyone observe similar p= roblem? Thanks. - Hui --_000_CA45F7026312hkangcsstonybrookedu_ Content-Type: text/html; charset="us-ascii" Content-ID: <7768C97B44E59841B58F18FB83D9F7F4@cs.stonybrook.edu> Content-Transfer-Encoding: quoted-printable

Hi,

I am running two hosts with xen-4.0.1 and dom0 kernel 2.6.32.41. The t= wo hosts are connected to a Gigbit switch. Then I use iperf to measure the = bandwidth and got the results of 200Mbit/second, which is only 1/5 of the a= vailable bandwidth.

Then I boot these two hosts using a vanilla kernel of 2.6.32.34. The i= perf reports 937Mbis/s, which is very close to the available link bandwidth= .

So why the pvops dom0 kernel performs so bad? Does anyone observe simi= lar problem? Thanks.

- Hui

--_000_CA45F7026312hkangcsstonybrookedu_-- --===============0386370709== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel --===============0386370709==-- From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: dom0 kernel v2.6.32.41 network problem Date: Fri, 15 Jul 2011 14:19:33 -0400 Message-ID: <20110715181933.GA18907@dumpdata.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Hui Kang Cc: "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org On Fri, Jul 15, 2011 at 05:54:11PM +0000, Hui Kang wrote: > Hi, > I am running two hosts with xen-4.0.1 and dom0 kernel 2.6.32.41. The two hosts are connected to a Gigbit switch. Then I use iperf to measure the bandwidth and got the results of 200Mbit/second, which is only 1/5 of the available bandwidth. > Then I boot these two hosts using a vanilla kernel of 2.6.32.34. The iperf reports 937Mbis/s, which is very close to the available link bandwidth. > So why the pvops dom0 kernel performs so bad? Does anyone observe similar problem? Thanks. Weird. No idea - I get around 911Mb/s with my dom0 - which is 3.0.0-rc7. Do you have some debug options turned on? > > - Hui > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Campbell Subject: Re: dom0 kernel v2.6.32.41 network problem Date: Fri, 15 Jul 2011 19:36:41 +0100 Message-ID: <1310755001.11556.14.camel@dagon.hellion.org.uk> References: <20110715181933.GA18907@dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20110715181933.GA18907@dumpdata.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Konrad Rzeszutek Wilk Cc: "xen-devel@lists.xensource.com" , Hui Kang List-Id: xen-devel@lists.xenproject.org On Fri, 2011-07-15 at 19:19 +0100, Konrad Rzeszutek Wilk wrote: > On Fri, Jul 15, 2011 at 05:54:11PM +0000, Hui Kang wrote: > > Hi, > > I am running two hosts with xen-4.0.1 and dom0 kernel 2.6.32.41. The two hosts are connected to a Gigbit switch. Then I use iperf to measure the bandwidth and got the results of 200Mbit/second, which is only 1/5 of the available bandwidth. > > Then I boot these two hosts using a vanilla kernel of 2.6.32.34. The iperf reports 937Mbis/s, which is very close to the available link bandwidth. > > So why the pvops dom0 kernel performs so bad? Does anyone observe similar problem? Thanks. It is definitely worth checking that you don't have any debug options enabled in your kernel .config. Some of those options are far more expensive under Xen than native. (unfortunately I'm drawing a blank on precisely which options we've seen this with right now, anything to do with debugging page allocations etc would be a good bet). Ian.