<div dir="ltr">Hi James,<div><br></div><div>Let me ask you something...</div><div><br></div><div>Are you using the package `openvswitch-datapath-dkms' from Havana Ubuntu Cloud Archive with Linux 3.8?</div><div><br></div>
<div>I am unable to compile that module on top of Ubuntu 12.04.3 (with Linux 3.8) and I'm wondering if it is still required or not...</div><div><br></div><div>Thanks!</div><div>Thiago</div></div><div class="gmail_extra">
<br><br><div class="gmail_quote">On 2 October 2013 06:14, James Page <span dir="ltr"><<a href="mailto:james.page@ubuntu.com" target="_blank">james.page@ubuntu.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
-----BEGIN PGP SIGNED MESSAGE-----<br>
Hash: SHA256<br>
<br>
Hi Folks<br>
<br>
I'm seeing an odd direction performance issue with my Havana test rig<br>
which I'm struggling to debug; details:<br>
<br>
Ubuntu 12.04 with Linux 3.8 backports kernel, Havana Cloud Archive<br>
(currently Havana b3, OpenvSwitch 1.10.2), OpenvSwitch plugin with GRE<br>
overlay networks.<br>
<br>
I've configured the MTU's on all of the physical host network<br>
interfaces to 1546 to add capacity for the GRE network headers.<br>
<br>
Performance between instances within a single tenant network on<br>
different physical hosts is as I would expect (near 1GBps), but I see<br>
issues when data transits the Neutron L3 gateway - in the example<br>
below churel is a physical host on the same network as the layer 3<br>
gateway:<br>
<br>
ubuntu@churel:~$ scp hardware.dump <a href="http://10.98.191.103" target="_blank">10.98.191.103</a>:<br>
hardware.dump<br>
100% 67MB 4.8MB/s<br>
00:14<br>
<br>
ubuntu@churel:~$ scp 10.98.191.103:hardware.dump .<br>
hardware.dump<br>
100% 67MB<br>
66.8MB/s 00:01<br>
<br>
As you can see, pushing data to the instance (via a floating ip<br>
10.98.191.103) is painfully slow, whereas pulling the same data is<br>
x10+ faster (and closer to what I would expect).<br>
<br>
iperf confirms the same:<br>
<br>
ubuntu@churel:~$ iperf -c 10.98.191.103 -m<br>
- ------------------------------------------------------------<br>
Client connecting to 10.98.191.103, TCP port 5001<br>
TCP window size: 22.9 KByte (default)<br>
- ------------------------------------------------------------<br>
[ 3] local 10.98.191.11 port 55330 connected with 10.98.191.103 port 5001<br>
[ ID] Interval Transfer Bandwidth<br>
[ 3] 0.0-10.0 sec 60.8 MBytes 50.8 Mbits/sec<br>
[ 3] MSS size 1448 bytes (MTU 1500 bytes, ethernet)<br>
<br>
ubuntu@james-page-bastion:~$ iperf -c 10.98.191.11 -m<br>
<br>
<br>
- ------------------------------------------------------------<br>
Client connecting to 10.98.191.11, TCP port 5001<br>
TCP window size: 23.3 KByte (default)<br>
- ------------------------------------------------------------<br>
[ 3] local 10.5.0.2 port 52190 connected with 10.98.191.11 port 5001<br>
[ ID] Interval Transfer Bandwidth<br>
[ 3] 0.0-10.0 sec 1.07 GBytes 918 Mbits/sec<br>
[ 3] MSS size 1448 bytes (MTU 1500 bytes, ethernet)<br>
<br>
<br>
918Mbit vs 50Mbits.<br>
<br>
I tcpdump'ed the traffic and I see alot of duplicate acks which makes<br>
me suspect some sort of packet fragmentation but its got me puzzled.<br>
<br>
Anyone have any ideas about how to debug this further? or has anyone<br>
seen anything like this before?<br>
<br>
Cheers<br>
<br>
James<br>
<br>
<br>
- --<br>
James Page<br>
Ubuntu and Debian Developer<br>
<a href="mailto:james.page@ubuntu.com">james.page@ubuntu.com</a><br>
<a href="mailto:jamespage@debian.org">jamespage@debian.org</a><br>
-----BEGIN PGP SIGNATURE-----<br>
Version: GnuPG v1.4.14 (GNU/Linux)<br>
Comment: Using GnuPG with Thunderbird - <a href="http://www.enigmail.net/" target="_blank">http://www.enigmail.net/</a><br>
<br>
iQIcBAEBCAAGBQJSS+QSAAoJEL/srsug59jD8ZcQAKbZDVU8KKa7hsic7+ulqWQQ<br>
EFbq8Im5x4mQY7htIvIOM26BR0ktAO5luE7zMBXsA4AwPud1BQSGhw89/NvNhADT<br>
TLcGdQADsomeiBpJebzwUmvL/tYUoMDRA3O96mUn2pi0fySWbEuEgMDjDJ/ow23D<br>
Y7nEv0mItaZ4MBSI9RZcqsDUl7UbbdlGejSWhJcwp/127HMU9nYwWNz5UHJjsGZ1<br>
eITyv1WZH/dYPQ1SES41qD1WvkTBugopGJvptEyrcO62A+akGOvnqpsHgPECbLb+<br>
b/8rk8nB1HB74Wh+tQP4WRQCZYso15nB6ukIyIU24Qti2tXtXDdKwszEoblCwCT3<br>
YZJTERNOENURlUEFwgi6FNL+nZomSG0UJU6qqDGiUJkbSF7SwJm4y8/XRlJM2Ihn<br>
wyxFB0qe3YdMqgDLZn11GwCDqn3g11hYaocHNUyRaj/tgxhGKbOFvix5kz3I4V7T<br>
gd+sqUySMVd9wCRXBzDDhCuG9xf/QY2ZQxXzyfPJWd9svPh/O6osTSQzaI1eZl9/<br>
jVRejMAFr6Rl11GPKd3DYi32GXa896QELjBmJ9Kof0NDlCcDuUKpVeifIhcbQZZV<br>
sWyQmbb6Z/ypFV9xXiLRfH2fW2bAQQHgiQGvy9apoE78BWYdnsD8Q3Ekwag6lFqp<br>
yUwt/RcRXS1PbLG4EGFW<br>
=HTvW<br>
-----END PGP SIGNATURE-----<br>
<br>
_______________________________________________<br>
Mailing list: <a href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack" target="_blank">http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack</a><br>
Post to : <a href="mailto:openstack@lists.openstack.org">openstack@lists.openstack.org</a><br>
Unsubscribe : <a href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack" target="_blank">http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack</a><br>
</blockquote></div><br></div>