08:30:05 <anac1> #startmeeting Yardstick work meeting
08:30:05 <collabot`> Meeting started Thu Apr  7 08:30:05 2016 UTC.  The chair is anac1. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:30:05 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:30:05 <collabot`> The meeting name has been set to 'yardstick_work_meeting'
08:30:14 <anac1> #info Ana Cunha
08:30:17 <mtahhan> Hi Folksis this an IRC only meeting?
08:30:24 <anac1> yes, IRC only
08:30:28 <mtahhan> cool :)
08:30:34 <mtahhan> #info Maryam Tahhan
08:30:39 <qiliang27> #info Qiliang
08:30:42 <patrick1> #info patrick
08:30:43 <rex_lee> #info Mingjiang
08:30:46 <kubi> #info kubi
08:31:01 <anac1> #topic new committers
08:31:23 <anac1> #info congratulations to kubi and vincenzo, new committers in yardstick !
08:31:35 <kubi> thank you~
08:31:39 <qiliang27> congratulations!
08:31:44 <rex_lee> congratuations!
08:31:46 <patrick1> congrats!
08:31:58 <anac1> #topic integration yardstick - vsperf
08:32:26 <anac1> mtahhan, ptl for vsperf, is attending today to sort out together the details of what needs to be done
08:32:52 <anac1> #link https://etherpad.opnfv.org/p/yardstick_release_c
08:33:02 <mtahhan> yeah, that and to address any questions that are still outstanding from the email I sent out
08:33:14 <andrasb> #info andrasb
08:33:25 <anac1> #info questions were answered by mtahhan on email to mailing list
08:33:36 <anac1> outstanding questions ?
08:34:06 <anac1> mtahhan: what vsperf test cases are intended for the integration ?
08:34:44 <mtahhan> #info RFC 2544 x% frameloss test – binary search
08:34:52 <mtahhan> #info RFC2544 Back to Back test
08:35:04 <mtahhan> #info Continuous test – blast the DUT at x% TX rate for y seconds and measure throughput/forwarding rate without regard for packet loss.
08:35:16 <mtahhan> #info Flow Scalablility RFC2544 test
08:35:26 <mtahhan> In the case of Moongen only a subset of these would be available. For HW traffic gens all would be available
08:35:53 <anac1> #info the integration will include only SW traffic gens, is that correct ?
08:36:58 <mtahhan> I was thinking about this a lot this week and I think we can integrate under yardstick as an option. We have a fixed POD setup in the OPNFV LAB that's connected to IXIA that could allow us to run yardstick tests with the HW traffic gens.
08:37:34 <mtahhan> even with the sw traffic gen we have a dependency on the existence of a physical NIC to do the timestamping of the packets
08:38:12 <mtahhan> and with the SW traffic gen only RFC 2544 x% frameloss test – binary search and Continuous test would be supported
08:38:24 <anac1> mtahhan: ok, but could we have 2 options: 1) only SW traffic gens (no dependency) 2) HW+SW traffic gens - does that make any sense ?
08:38:39 <mtahhan> we could
08:39:11 <mtahhan> it's just a configuration parameter into vsperf on the CLI - so it's doable
08:39:15 <anac1> so we can run on more PODs and a subset will have HW dependency to run additional tests
08:39:34 <mtahhan> that sounds good to me
08:39:47 <anac1> what is the opinons ? jnon, kubi, qiliang ?
08:40:16 <anac1> can we work on similar design as for apexlake?
08:42:33 <mtahhan> I think if I am correct it just uses DPDK pktgen at the traffic generator
08:43:00 <mtahhan> we could plug in at the point where pktgen is used
08:43:53 <kubi> anac1, i think so, vsperf has their own framework. and yardstick is a umbrella
08:45:10 <anac1> kubi: yes, so we need a scenario for vsperf ?
08:46:46 <anac1> so we need a scenario for vsperf, to run their test cases from yardstick, agree ?
08:47:11 <jnon> ye we should have a vsperf scenario
08:48:09 <kubi> i agree
08:48:12 <anac1> mtahhan: can you check if someone in vsperf could write a yardstick scenario for vsperf ? yardstick can support.
08:48:18 <jnon> should we add  DPDK pktgen  to yardstick image ?
08:48:58 <qiliang27> mtahhan: how is vsperf test case runned?
08:49:02 <anac1> jnon: i think vsperf needs DPDK pktgen, we have a new test case for DPDKpktgen
08:50:12 <mtahhan> I will check if we can write up the scenario yes.
08:50:32 <anac1> mtahhan: thanks
08:51:35 <mtahhan> well for our integration with Yardstick vsperf will run in traffic-gen only mode. which means that we will rely on yardstick to configure the VMs and vSwitches... we just configure the traffic gen, ask it to start the test and wait for the results to be sent back
08:51:47 <anac1> #info 2 options proposed: 1) vsperf test cases using SW traffic gen and 2) vsperf test cases using HW+SW traffic gen
08:52:22 <anac1> #info investigation is needed for understanding how the vsperf test cases are run
08:53:09 <jnon> how should the vswitches be configured? That will be done by the scenario right?
08:53:26 <mtahhan> Yes, I hope that will be covered by the scenario
08:53:42 <anac1> #info from the investigation, details on what additional functionality is required from yardstick
08:54:01 <anac1> mtahhan: what's the best way for us to understand how your test cases are run today?
08:54:27 <mtahhan> we can do a readout to you folks early next week if that works.
08:54:30 <anac1> #info the scenario should cover vswitch config
08:54:48 <anac1> mtahhan: thanks, what is the best day+time for you ?
08:55:30 <kubi> mtahhan:  could vsperf boot VMs and vSwitches with their own framework now?
08:55:43 <mtahhan> this time Tuesday would be perfect for an hour if that works
08:56:08 <mtahhan> yeah we can boot the vm and configure the vswitch in our framework today
08:56:20 <mtahhan> VMs**
08:56:44 <mtahhan> if not Tues than Mon but Tues would be my preference
08:56:45 <anac1> mtahhan: tuesday 12th April at 10:00 CEST works for you ?
08:56:45 <mtahhan> :D
08:57:13 <mtahhan> anac1 yeah that works great for me
08:57:39 <anac1> guys, tuesday 12th April at 10:00 CEST (UTC 8:00) ok ?
08:57:52 <anac1> 1 hour meeting, i fix gotomeeting
08:58:02 <jnon> i have other things on Tuesday
08:58:17 <jnon> Team building day :)
08:58:34 <anac1> kubi, qiliang ? tuesday ?
08:59:57 <kubi> i'm ok, but i think it will be better if we could find a time which is available for jnon
09:00:05 <anac1> mtahhan: i'll check with more team members, we say tuesday 12th preliminary, i'll book the meeting latest tomorrow
09:00:21 <anac1> i'll discuss with jnon
09:00:40 <anac1> would be good to have vincenzo also
09:00:44 <mtahhan> perfect
09:00:50 <mtahhan> thanks
09:01:11 <anac1> #action anac1 to agree with the team on a date + time and inform all latest tomotroow friday 8th
09:01:22 <anac1> mtahhan: thnaks for participating today
09:01:30 <anac1> thnaks everyone
09:01:36 <anac1> #endmeeting