<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
</head>
<body dir="ltr">
<div id="divtagdefaultwrapper" style="font-size:12pt;color:#000000;background-color:#FFFFFF;font-family:Calibri,Arial,Helvetica,sans-serif;">
<p></p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">Hi,</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"><br>
</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> I'm generally sympathetic to what you're saying, and I agree that we</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> need to do something about disabled-by-default features, at the very</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> least on the testing front. Comments in-line.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> On Mon, Feb 8, 2016 at 4:47 PM, Sean M. Collins <sean at coreitpro.com> wrote:</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Hi,</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > With the path_mtu issue - our default was to set path_mtu to zero, and</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > do no calculation against the physical segment MTU and the overhead for</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > the tunneling protocol that was selected for a tenant network. Which</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > means the network would break.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > I am working on patches to change our behavior to set the MTU to 1500 by</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > default[1], so that at least our out of the box experience is more</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > sensible.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > This brings me to the csum feature of recent linux kernel versions and</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > related network components.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Patches:</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > https://review.openstack.org/#/c/220744/</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > https://review.openstack.org/#/c/261409/</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Bugs/RFEs:</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > https://bugs.launchpad.net/neutron/+bug/1515069</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > https://bugs.launchpad.net/neutron/+bug/1492111</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Basically, we see that enabling the csum feature creates the conditions</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > where 10gig link were able to be fully utilized[2] in one instance[3]. My</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > thinking is - yes I too would like to fully utilize the links that I</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > paid good money for. Someone with more knowledge can correct me</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > , but is there any reason not to enable this feature? If your hardware</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > supports it, we should utilize it. If your hardware doesn't support it,</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > then we shouldn't.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > tl;dr - why do we keep merging features that create more knobs that</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > deployers and deployment tools need to keep turning? The fact that we</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > merge features that are disabled by default means that they are not as</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > thoroughly tested as features that are enabled by default.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> That is partially a testing issue which fullstack is supposed to</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> solve. We can't afford to set up a job for every combination of</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> Neutron configuration values, not upstream and not in different</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> downstream CI environments. Fullstack can test different</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> configurations knobs quickly, and it's something that a developer can</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> do on his own without depending on infra changes. It's also easy to</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> run, and thus easy to iterate.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> As for concrete actions, we do have a fullstack test that enables</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> l2pop and checks connectivity between two VMs on different nodes. It's</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> the only code patch that actually covers l2pop at the upstream gate!</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> It already caught a regression that Armando and I fixed a while ago.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> As for DVR, I'm searching for someone to pick up the gauntlet and</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> contribute some L3 fullstack tests. I'd be more than happy to review</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> it! I even have an abandoned patch that gets the ball rolling (The</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> idea is to test L3 east/west, north/south with FIP and north/south</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> without FIP for all four router types: Legacy, HA, DVR and DVR HA. You</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> run the same test in four different configurations, fullstack is</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> basically purpose built for this).</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"></p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">I am adding something similar to Dragonflow. It creates tap devices that mimic VMs connected via interfaces to the switch. This gives us complete control inside the tests to send and received
a predefined set of packets though simple python code written inside the test.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"><br>
</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">I have successfully written an ARP responder test, and other tests (such as DHCP server testing, ping echo/reply tests, and the such) are in the works.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"><br>
</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">Currently, this set-up is geared towards a single compute-node testing. It can be easily extended further, to multi-node tests, incl. east-west communication, and north-south communication with
and without FIP. It can also be used to verify that all the flags on the packet are set correctly.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"><br>
</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">Currently, these tests are in review in Dragonflow[4], but all of the heavy lifting is done either done via the Neutron API, or directly with the OVS switch. We could easily integrate it with
neutron tests.</p>
<br>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">[4] https://review.openstack.org/275714</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;"><br>
</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Neutron should have a lot of things enabled by default that improve</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > performance (l2pop? path_mtu? dvr?), and by itself, try and enable these</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > features. If for some reason the hardware doesn't support it, log that</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > it wasn't successful and then disable.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> I don't know if this is what you wanted to talk about (It feels more</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> like a side note to me, so I'm sorry if I'm about to hijack the</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> conversation!), but I think that if an admin sets a certain</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> configuration option, the software should respect it in a predictable</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> manner. If an agent tries to use a certain config knob and fails, it</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> should error out (Saying specifically what's wrong), and not disable</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> the option but keep on living, because that is surprising behavior,</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> and there's nothing telling the admin that the option he expects to be</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> on is actually off, until he notices it the hard way some time later.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > OK - that's it for me. Thanks for reading. I'll put on my asbestos</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > undies now.</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > [1]: https://review.openstack.org/#/c/276411/</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > [2]: http://openvswitch.org/pipermail/dev/2015-August/059335.html</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > [3]: Yes, it's only one data point....</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > --</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Sean M. Collins</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > </p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > __________________________________________________________________________</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > OpenStack Development Mailing List (not for usage questions)</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > Unsubscribe: OpenStack-dev-request at</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > lists.openstack.org?subject:unsubscribe</p>
<p style="margin: 0px; text-indent: 0px; white-space: pre-wrap;">> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev</p>
<br>
<p></p>
</div>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span style="font-size:7.0pt;font-family:"Arial","sans-serif";color:gray">-------------------------------------------------------------------------------------------------------------------------------------------------</span></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;line-height:12.0pt">
<span style="font-size:7.0pt;font-family:"Arial","sans-serif";color:gray">This email and any files transmitted and/or attachments with it are confidential and proprietary information of<br>
Toga Networks Ltd., and intended solely for the use of the individual or entity to whom they are addressed.<br>
If you have received this email in error please notify the system manager. This message contains confidential<br>
information of Toga Networks Ltd., and is intended only for the individual named. If you are not the named<br>
addressee you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately<br>
by e-mail if you have received this e-mail by mistake and delete this e-mail from your system. If you are not
<br>
the intended recipient you are notified that disclosing, copying, distributing or taking any action in reliance on<br>
the contents of this information is strictly prohibited. </p>
</span>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span style="font-size:7.0pt;font-family:"Arial","sans-serif";color:gray">------------------------------------------------------------------------------------------------------------------------------------------------</span></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;line-height:12.0pt">
<span style="font-size:7.0pt;font-family:"Arial","sans-serif";color:gray"><br>
</p>
</span>
</body>
</html>