08:31:15 <rbbratta> #startmeeting yardstick work meeting
08:31:15 <collabot> Meeting started Mon Oct  9 08:31:15 2017 UTC.  The chair is rbbratta. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:31:15 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:31:15 <collabot> The meeting name has been set to 'yardstick_work_meeting'
08:31:22 <rbbratta> #topic roll call
08:31:31 <JingLu> #info JingLu
08:32:22 <rbbratta> not many people here.  Must have been a good holiday. :)
08:32:40 <JingLu> LOL
08:32:41 <Mingjiang> #info Rex
08:32:58 <JackChan> #info Jack
08:33:35 <rbbratta> How was the Holiday?  Everyone had a good time?
08:34:31 <rbbratta> #topic Euphrates
08:34:43 <rbbratta> new release date is Oct. 20th
08:35:09 <JingLu> Got it
08:35:22 <rbbratta> #info Fuel is working now, but they are hitting Ping test issue  https://jira.opnfv.org/browse/YARDSTICK-809
08:35:27 <Mingjiang> great holiday and return to work today, just in time for the meeting
08:36:01 <rbbratta> Mingjiang:  :)
08:36:16 <rbbratta> so we have more time to fix issues in Euphrates.
08:36:40 <rbbratta> I haven't checked Compass status.
08:37:40 <JingLu> compass is only working in virtual  pods so far
08:37:49 <kubi001> #info Kubi
08:38:51 <rbbratta> For https://jira.opnfv.org/browse/YARDSTICK-809 Ping test, did we have a plan for a fix?
08:39:30 <JackChan> rbbratta: not sure, have we come to a conclusion?
08:40:37 <JingLu> The compass pods are offline during the Chinese national holiday , so CI jobs have been suspended.
08:42:35 <Mingjiang> JingLu: do compass physical pods available now? when it will be available?
08:42:50 <abhijitsinha> #info Abhijit
08:43:13 <rbbratta> JackChan: ping fails so we raise RuntimeError?
08:43:27 <JingLu> Mingjiang no compass job is running now
08:43:45 <rbbratta> JackChan: I'll try to look at it tomorrow
08:44:07 <JingLu> Mingjiang for stable branch.
08:44:28 <JackChan> rbbratta: I don't think so. Just sometimes ping failed, last meeting we plan to give it a default value.
08:45:23 <JackChan> rbbratta: for example: sla or 2 * sla.
08:46:48 <rbbratta> JackChan: okay I checked the meeting logs.  But the lost pings aren't considered a failure
08:47:59 <rbbratta> Okay, so in general everyone is working on Euphrates bugs and checking CI results.
08:48:09 <JackChan> rbbratta: yes, actually it can be pinged, just timeout sometimes
08:48:45 <rbbratta> JackChan: so there is packet loss?
08:49:00 <JackChan> rbbratta: yes
08:49:32 <rbbratta> JackChan: and packet loss > 0 isn't considered a failure?
08:50:02 <rbbratta> #topic task queue change
08:50:59 <rbbratta> #info https://gerrit.opnfv.org/gerrit/#/c/44147/  changed how we drain the result queue.
08:51:19 <rbbratta> We had a few tests hang in the duration and iteration runners.
08:51:34 <JackChan> rbbratta: not sure actually, still need to be discussed.
08:51:48 <rbbratta> it turns out that output_queue and result_queue were not being read until the test finished.
08:52:23 <rbbratta> so I added a loop in base.Runner.join() to read from the queue while the test is running.
08:52:51 <rbbratta> I'm surprised no other test failed, we should have been having problems just filling the queues to max size.
08:53:43 <rbbratta> I also changed the queue get and put operations to be non-blocking.
08:53:51 <JackChan> rbbratta: so we want read result from queue in real time, right?
08:55:01 <rbbratta> yes while the runners are running we will read the queue and save it in runner.  So the queues should be empty most of the time.
08:55:37 <rbbratta> we could in the future add another loop to push the results to influxdb during the test, but that is probably not for Euphrates.
08:55:50 <JackChan> rbbratta: OK, that's what I want to do, thanks
08:57:15 <rbbratta> the main runners have been changed, and it doesn't look like anything has broken.    just be mindful of blocking queue operations
08:57:36 <rbbratta> and we have more time to check for regressions in Euphrates results.
08:57:51 <rbbratta> anything else for Euphrates?
08:58:42 <rbbratta> There was some discussion about reporting Fuel x86 and arm results separately, but I haven't seen any patches
08:59:34 <JackChan> rbbratta: yes, I have fixed it in testresults.opnfv.org, but not a patch.
08:59:54 <rbbratta> JackChan: cool, thanks
09:00:18 <Mingjiang> rbbratta: what about eliminating 5.1, 5.2? i've received a mail from David about it
09:00:37 <rbbratta> Mingjiang: people are still voting I think
09:00:51 <JackChan> rbbratta: Since we do not have a way to distinguish *86 or arm from database.
09:00:58 <Mingjiang> what does it mean, i don't quite understand
09:01:24 <rbbratta> Mingjiang: after 5.0 on Oct 20th the next release will be Dec. 15th.
09:01:40 <rbbratta> that is the proposal
09:02:07 <rbbratta> only having one additional release for Euphrates on Dec. 15th,  which is after plugfest.
09:03:53 <rbbratta> hopefully there will be no delays for the Dec 15th release, CI should be working and no surprise holidays.  :)
09:04:02 <Mingjiang> got it, thanks
09:05:13 <rbbratta> I haven't voted yet, but I think the proposal will win.
09:05:28 <rbbratta> #topic AoB
09:06:29 <rbbratta> in a few weeks there will be more Intel developers attending this meeting from Ireland (Europe time zone).
09:06:57 <rbbratta> We might want to add Gotomeeting for this time slot.
09:07:06 <Mingjiang> nice to hear it
09:07:48 <rbbratta> are there any issues with using Gotomeeting for this time slot.  It does slow down the meeting sometimes.
09:08:50 <JingLu> I think it will not be a problem
09:08:59 <rbbratta> We can discuss in future meetings
09:09:49 <Mingjiang> ok
09:10:24 <rbbratta> anything else?
09:10:44 <rbbratta> It is good to have you all back, it has been lonely this past week.  :)
09:12:02 <kubi001> :)
09:12:27 <Mingjiang> glad to be back :)
09:13:33 <rbbratta> okay, thanks everyone, see you tomorrow
09:13:46 <rbbratta> #endmeeting