08:32:14 <kubi001> #startmeeting Yardstick Work Meeting
08:32:14 <collabot`> Meeting started Mon May 22 08:32:14 2017 UTC.  The chair is kubi001. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:32:14 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:32:14 <collabot`> The meeting name has been set to 'yardstick_work_meeting'
08:32:43 <kubi001> #topic roll call
08:32:44 <Mingjiang> #info Rex
08:32:50 <rbbratta> #info Ross
08:32:50 <JingLu> #info JingLu
08:32:56 <Kanglin> #info Kanglin
08:33:02 <JackChan> #info Jack
08:33:03 <abhijitsinha> #info Abhijit
08:33:14 <zhihui> #zhihui_wu
08:33:43 <zhihui> #info zhihui_wu
08:34:25 <kubi001> #topic action items follow up
08:34:35 <kubi001> #link http://ircbot.wl.linuxfoundation.org/meetings/opnfv-yardstick/2017/opnfv-yardstick.2017-05-15-08.30.html
08:35:23 <kubi001> #info AP1:  Rex will update gotomeeting link in the wiki before tomorrow --> Done
08:35:25 <kubi001> #link https://global.gotomeeting.com/join/391235029
08:36:31 <kubi001> #info AP2: Ross will try to update jira tomorrow
08:36:57 <kubi001> #info AP3: Kanglin will update his jira before tomorrow
08:37:13 <kubi001> #info we will check the JIRA at tomorrow meeting
08:37:52 <kubi001> #topic agenda
08:38:22 <kubi001> we didn't draft today's meeting agenda, so please feel free to propose
08:38:36 <kubi001> from my side, PTL election would be a topic
08:39:08 <kubi001> Beijing summit schedule
08:39:23 <kubi001> any topic from you?
08:39:46 <rbbratta> Jenkins hang from last week
08:39:51 <JackChan> I think we need to unify our output.
08:40:08 <kubi001> rbbratta: OK, good point
08:40:27 <kubi001> PTL election
08:40:32 <kubi001> Beijing summit schedule
08:40:39 <kubi001> Jenkins hang
08:40:59 <JackChan> Currently we have file, influxdb, mongodb. But our output format is not unified. They are all different.
08:41:21 <kubi001> output format issue
08:41:34 <kubi001> OK, let's move to PTL election
08:41:38 <kubi001> #topic PTL election
08:42:41 <kubi001> #info PTL nominaton have been end since last friday. the candidate is Ross rbbratta
08:44:19 <kubi001> #info As we only have one candidate, we needn't to vote with http://civs.cs.cornell.edu/
08:44:50 <kubi001> #info we will use the gerrit to vote directly
08:45:42 <kubi001> any comment about this vote process?
08:46:27 <rbbratta> sounds good
08:46:33 <JingLu> Agree
08:46:47 <JackChan> +1
08:46:50 <zhihui> +1
08:47:01 <kubi001> #action kubi will upload a patch to ask all committers to vote for new PTL
08:47:03 <Kanglin> +1
08:47:03 <Mingjiang> +1
08:47:14 <abhijitsinha> +1
08:47:17 <kubi001> OK. great
08:47:50 <kubi001> #info vote will be closed at this friday. please vote once the patch is available
08:48:02 <kubi001> #topic Beijing Summit
08:48:18 <kubi001> #link https://wiki.opnfv.org/display/EVNT/Beijing+Design+Summit+Schedule
08:48:49 <kubi001> #info I booked the 2 hours for yardstick session
08:48:54 <rbbratta> looks like my presentation is at 13:45 on Monday
08:50:17 <kubi001> looks good, yardstick session will be started at 15:30
08:51:28 <rbbratta> Do we need more Yardstick session on Day 2?
08:52:52 <kubi001> I'm not sure, It depends on what will be discussed
08:52:52 <rbbratta> Anyone looked at the schedule for the Wed-Thurs?
08:53:20 <Mingjiang> I thinks there's still time slot available
08:53:54 <zhihui> What will be on the agenda for Yardstick session?
08:54:20 <rbbratta> Jing Zhang has a presentation on Yardstick Wed 3:50pm
08:55:04 <kubi001> #link http://events.linuxfoundation.org/events/opnfv-summit/program/schedule
08:55:07 <rbbratta> We will probably go through the list of features for Euphrates again and drill down into specifics
08:55:44 <rbbratta> Obviously, I'd like to go over the NSB Heat design.
08:56:09 <kubi001> sounds good
08:56:48 <rbbratta> Do we want a separate OpenRetriever meeting at summit?
08:56:49 <zhihui> I am interested in NSB Heat design.
08:57:01 <Mingjiang> we can also talk about yardstick GUI or framework evolement
08:57:29 <kubi001> rbbratta: not yet, It is a good point.
08:57:57 <kubi001> we can talk with OR ptl
08:58:26 <kubi001> Mingjiang: yes, good idea for yardstick session
08:58:44 <JackChan> Mingjiang: I am also interested in GUI.
08:58:55 <rbbratta> so maybe we should take the list of features and estimate how much time to spend on them.
08:59:04 <kubi001> rbbratta:  we are trying to have a joint meeting with OR team
08:59:21 <rbbratta> kubi001: maybe on Day 2, in one of the available slots?
08:59:42 <kubi001> OK,  will do
08:59:51 <zhihui> Can we list the topics in etherpad or wiki page?
09:00:01 <kubi001> for Day 2, we have a invitation from Doc team
09:00:13 <kubi001> Documentation / testing working group
09:00:25 <kubi001> 13:00 - 14:00 Day 2
09:00:30 <rbbratta> we have beijing summit etherpad I think
09:01:15 <kubi001> rbbratta: yes
09:01:15 <JingLu> #link https://etherpad.opnfv.org/p/yardstick_beijing_summit_planning
09:01:18 <kubi001> #link https://etherpad.opnfv.org/p/yardstick_beijing_summit_planning
09:01:24 <kubi001> JingLu: thanks
09:01:46 <JingLu> :D
09:02:03 <kubi001> #action please update your idea and list the topic which you want to discuss at design summit session
09:02:13 <zhihui> Maybe it is better to add some time infomation. It will be nice for people who cannot attend all the topics.
09:02:28 <kubi001> zhihui: make sense
09:02:52 <kubi001> #info t is better to add some time infomation
09:02:57 <kubi001> #undo
09:02:57 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x20cd910>
09:02:58 <rbbratta> yes, the big features might take at least an hour.
09:03:03 <kubi001> #info it is better to add some time infomation
09:03:15 <kubi001> Agree
09:03:28 <kubi001> OK. next topic?
09:03:48 <kubi001> #topic jenkins hang issue
09:04:10 <rbbratta> I was going to write up a root cause, but basically the SSH Mock broke
09:04:52 <rbbratta> When I refactored from SSH() to SSH.from_node() I thought I fixed all the unittests, but the gerrit change was not merged, so it was never updated.
09:05:51 <kubi001> JingLu have the right to cancel  jenkins job
09:06:00 <rbbratta> If we had rebased the MultiMonitor patch then it would have failed verification.
09:06:37 <rbbratta> So I think we need to rebase gerrit reviews more frequently, so we get up to date verified +1 from Jenkins.
09:06:50 <rbbratta> but the each time we rebase, we lose the +1, +2 CR.
09:07:42 <rbbratta> And the gerrit 'recheck' command doesn't do a rebase, so it isn't really useful.
09:08:02 <kubi001> recheck means reverify
09:08:25 <rbbratta> yes, but it just re-runs with the same old code.
09:09:22 <kubi001> yes
09:09:48 <rbbratta> we really want to re-verify by rebasing on top of latest master, in case master changed something.
09:09:58 <Mingjiang> I think, as commiters, we need to review more often and remember to rebase if a patch is not quite up-to-date.
09:10:10 <kubi001> agree
09:10:16 <rbbratta> agree
09:10:35 <rbbratta> There is a gerrit setting that can be enabled that keeps +1 CR on "trivial rebase"
09:11:06 <rbbratta> copyMinScore = true
09:11:06 <rbbratta> copyAllScoresOnTrivialRebase = true
09:11:06 <rbbratta> copyAllScoresIfNoCodeChange = true
09:11:36 <JingLu> rbbratta Where to set this?
09:11:53 <rbbratta> project.config  [label "Code-Review"]
09:12:30 <JackChan> rbbratta: in releng code?
09:12:38 <rbbratta> you have to manually change the gerrit settings
09:13:12 <rbbratta> git fetch origin refs/meta/config:refs/remotes/origin/meta/config
09:13:40 <rbbratta> so we should probably ask helpdesk
09:14:12 <zhihui> I guess the jenkins modification should be in releng, cannot be manually modified.
09:15:21 <rbbratta> the Jenkins timeout fix was in releng.  https://gerrit.opnfv.org/gerrit/#/c/34997/2
09:16:04 <rbbratta> The main point is to rebase more often.  We can follow up on gerrit tweaks later
09:16:33 <kubi001> #info we should rebase more often
09:16:53 <JingLu> not sure if we should manually change the gerrit settings, better ask helpdesk
09:17:08 <rbbratta> yes, we can ask helpdesk.
09:18:28 <kubi001> rbbratta: could you help to trace this issue with helpdesk?
09:19:16 <rbbratta> yes, I will write an email.
09:19:17 <kubi001> #topic output format issue
09:19:24 <kubi001> #undo
09:19:24 <collabot`> Removing item from minutes: <MeetBot.ircmeeting.items.Topic object at 0x20dded0>
09:20:02 <JackChan> I write a demo.
09:20:11 <kubi001> #action rbbratta will write an email to helpdesk to trace the configure for rebase
09:20:18 <kubi001> #topic  output format issue
09:20:25 <JackChan> #link http://paste.openstack.org/show/610125/
09:21:01 <JackChan> This a result format demo. Please feel free to comment.
09:23:06 <JackChan> Currently yardstick result format is not easy to use. We don't have a criteria field for every task, also we do not have it of each testcase. This is hard to use.
09:23:59 <JackChan> Also the result format store in file, influxdb, mongodb is different.
09:24:12 <JackChan> I think we need to unify them.
09:25:57 <kubi001> JackChan: I'm not very clear about your idea
09:28:15 <Mingjiang> what's the difference?
09:28:55 <Mingjiang> can you introduce how you add? or can you share what's it's like before?
09:29:01 <JackChan> Now we have three output ways: file, influxdb, mongodb. And their format is different. If a user want to get a 'status' field, he can get it from file, but he can't get this field in influxdb or mongodb.
09:30:13 <JackChan> now in influxdb and mongodb we do not have any field to show which test case is pass and which test case is fail.
09:32:20 <Mingjiang> so the criteria you added is to show if the testcase is Pass or Fail, like functest?
09:32:21 <JackChan> So I plan to unify them and make them have the same format. And add criteria field for every task and each test case.
09:32:31 <rbbratta> we should have consistent schema and convert it to the equivalent output.
09:32:49 <rbbratta> do we have enough info to derive a common schema?
09:32:53 <JackChan> rbbratta: yes
09:34:55 <JackChan> rbbratta: what information?
09:35:46 <rbbratta> do we know what fields we want in the output?
09:36:24 <JackChan> I think this is what we should discuss.
09:37:34 <JackChan> What I show is just a demo. And I am waiting for more good suggestions.
09:38:12 <kubi001> JackChan: sounds good.
09:38:36 <kubi001> As time is late, maybe we can review and comments after the meeting, and discuss at next meeting
09:38:43 <Mingjiang> I think you can prepare a patch and we can discuss more tomorrow. what you're doing looks like need a lot change.
09:39:25 <kubi001> Mingjiang: Agree
09:39:38 <JackChan> Mingjiang: yes, we need change a lot.
09:39:55 <kubi001> OK. It is time to end the meeting
09:40:09 <kubi001> thanks all for today
09:40:13 <kubi001> #endmeeting