14:59:23 <parus> #startmeeting massively_distributed_clouds
14:59:24 <openstack> Meeting started Wed Aug  2 14:59:23 2017 UTC and is due to finish in 60 minutes.  The chair is parus. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:59:27 <openstack> The meeting name has been set to 'massively_distributed_clouds'
14:59:54 <jamemcc> Hello
15:00:03 <parus> Hi Jamecc
15:00:05 <dpertin> Hi o/
15:00:24 <samP> Hi o/
15:00:27 <parus> Let's give it a minute for people to join.
15:01:01 <ansmith> o/
15:01:24 <kgiusti> o/
15:01:44 <jamemcc> samP has joined for the topic of LCOO-Extreme Testing initiative and any overlap to our work here and possibly ENOS
15:01:58 <parus> Let's get started
15:02:08 <samP> jamemcc: thx
15:02:32 <jamemcc> He had joined 2 meetings ago at my request but we didn't get to him.  Past midnight for Sampath.  Thank you Sampath.
15:02:52 <parus> Good.
15:03:10 <parus> Adrien is on vacation today..... so I will be chairing.
15:03:21 <jamemcc> Thank you
15:03:27 <parus> we can put samP early in the agenda.
15:03:37 <parus> Is there any other request for the agenda?
15:04:01 <parus> Let's start with announcements.
15:04:10 <parus> #topic announcements
15:04:46 <parus> About opendev. Afew of us are confirmed.
15:05:05 <parus> Please put your name on the etherpad if you would be attending.
15:05:23 <dpertin> #link http://www.opendevconf.com/
15:05:23 <knikolla> o/
15:05:43 <parus> We are trying to have a face to face meeting for those will be attending on september 6th.
15:06:18 <parus> Moving on.
15:06:41 <parus> There was a note from Adrien about new people from Ericsson wanting to participate.
15:07:13 <parus> If there are any new members in attendance today, this is a good time to introduce yourself.
15:08:03 <parus> Let's move to first topic.
15:08:34 <parus> #topic LCOO-Extreme
15:08:52 <samP> Hi, I'm Sampath from NTT
15:09:00 <samP> #link https://specs.openstack.org/openstack/openstack-user-stories/user-stories/proposed/openstack-extreme-testing.html
15:09:18 <samP> ^^ This is the userstory for extreme testing.
15:09:41 <samP> There I propose to add integration testing, component interface testing,
15:09:47 <samP> operational acceptance testing, destructive testing, concurrent testing,
15:09:57 <samP> performance testing, etc..
15:10:04 <samP> As a part of our work, we started from destructive testing.
15:10:28 <samP> Here is the link for QA spec of destructive testing,
15:10:34 <samP> #link https://review.openstack.org/#/c/443504/
15:11:07 <parus> Is Massively Distributed part of the scope?
15:11:42 <samP> parus: At the time, it was not
15:12:03 <parus> Are you actually doing the testing, or just specifying what tests to run?
15:13:02 <samP> parus: Now we are doing it by our self. Goal of this proposal is to propose common framework for those testing
15:14:18 <parus> I can see this is interesting for this team.
15:14:29 <parus> Do you have any suggestions on what next steps would be?
15:15:07 <samP> parus: we are now discussing about test architecture and other sub topics
15:15:46 <samP> parus: let me find the link...
15:16:03 <parus> samP: From a FEMDC perspectivem I believe there are slightly different tests that could be run.
15:16:47 <parus> samP: destructive is part of the scope... but we are really at the stage where we are trying to understand what the architecture options are that willl enable to scale well.
15:16:51 <samP> #link Task List and current Status https://openstack-lcoo.atlassian.net/wiki/spaces/LCOO/pages/13393034/Working+Space+for+Exteme+Testing+and+related+topics
15:17:15 <samP> #link https://openstack-lcoo.atlassian.net/wiki/spaces/LCOO/pages/15477787
15:18:06 <samP> parus: got it. we can include that POV as well.
15:18:57 <parus> samP: is there representation from this team in your meetings?
15:19:58 <samP> parus: NO,currently only 3 of us in VoIP
15:20:32 <jamemcc> Next team meeting is in Mexico City at the Ops Meetup
15:20:41 <samP> jamemcc: thanks
15:20:52 <jamemcc> But the subsequent week 8/16 there shoudl be a VOIP call
15:20:58 <samP> and we proposed a session in Ops Meetup at Mexico
15:21:51 <samP> #link https://etherpad.openstack.org/p/MEX-ops-meetup
15:21:56 <parus> Is there anyone from this team attending the Ops Meetup?
15:23:00 <parus> jamecc: can you be the link between the groups?
15:23:29 <jamemcc> Yes
15:23:32 <samP> I can announce our next VoIP details to your ML  [FEMDC]
15:23:44 <samP> jamemcc: sorry.. go ahead
15:23:49 <parus> Great. thanks
15:23:56 <jamemcc> Both are good
15:24:06 <parus> #action samP to announce next meeting
15:24:21 <samP> parus: jamemcc: sure, thanks
15:24:26 <parus> #action jamecc to act as link between groups.
15:24:38 <parus> LEt's move to next topic.
15:24:48 <parus> #topic AMQP alternatives
15:24:50 <samP> parus: thanks
15:25:01 <parus> Is there any update?
15:25:08 <kgiusti> Hi - I've been working on a critical oslo bug
15:25:20 <kgiusti> that affects alternative backends.
15:25:29 <kgiusti> basically breaks nova cells
15:25:52 <kgiusti> Rabbitmq is not affected, but all the others (zeromq, amqp1, kafka) are
15:26:07 <kgiusti> I think I have a fix for amqp1 - in progress.
15:26:24 <kgiusti> But...  I have no idea how to fix this for the remaining backends
15:26:44 <kgiusti> We'll esp. need help for zeromq.
15:27:20 <kgiusti> Otherwise, ansmith is making progress on kolla ansible
15:27:30 <kgiusti> see epad line 1021
15:27:39 <kgiusti> thats all I have this week.
15:28:16 <parus> thanks for the update.
15:28:37 <msimonin1> one thing from my side
15:28:43 <msimonin1> regarding amqp alternatibe
15:28:52 <msimonin1> alternative
15:28:53 <msimonin1> *
15:29:32 <msimonin1> During last performance WG meeting I suggest to formalize the test plan and result in the performace team documentation
15:29:54 <msimonin1> It's a place that centralized a lot of perf. evaluation of OpenStack
15:30:07 <msimonin1> what do you think kgiusti, ansmith ?
15:30:32 <msimonin1> #link https://docs.openstack.org/developer/performance-docs/
15:30:52 <kgiusti> msimonin1: YES!!!
15:31:01 <ansmith> +1
15:31:16 <kgiusti> By all means these tests should be part of that.
15:31:27 <msimonin1> ok :)
15:31:53 <kgiusti> Otherwise I suspect they won't have the exposure to upstream projects that are focused on rabbitmq solely
15:32:20 <msimonin1> yes I agree
15:32:32 <kgiusti> So what was the reaction of the perf WG?
15:32:33 <parus> So what is the action?
15:33:07 <msimonin1> I can start to formalize except if kgiusti want to :)
15:34:04 <kgiusti> oop
15:34:05 <parus> #action msimonin1 to formalize test plan and resultsso it fits the format of the performance team
15:34:24 <msimonin1> parus: ack
15:34:29 <kgiusti> msimonin1: want to start an email thread about your thoughts on this?
15:34:50 <kgiusti> Unless you already have something to propose...
15:35:06 <parus> Let's take this to mailing list and  move to next topic
15:35:11 <kgiusti> +1
15:35:14 <msimonin1> +&
15:35:15 <msimonin1> +1
15:35:21 <parus> #topic cockroachdb
15:35:32 <parus> any update to discuss here?
15:35:43 <parus> we have 25 min left.
15:36:01 <msimonin1> no update from my side
15:36:13 <parus> Let's move to use cases then
15:36:19 <msimonin1> rcherrueau is on vacation let's wait for him
15:36:29 <parus> #topic UseCase discussion
15:36:55 <dsantoro> We started to add some motivation why we prefer to have k8s on top of OpenStack in slide 15
15:37:41 <dsantoro> moreover I've just replied to a couple of comments in slides 16 and 8 from Paul-André
15:37:57 <parus> Thank you for that.
15:38:25 <dsantoro> We slightly modified the 3rd use-case. It will be about car tracking with plate recognition instead of people tracking via WiFi probes
15:39:00 <dsantoro> From our side next steps are i) to explore ENOS to deploy an OpenStack instance in our Lab in order to be able to test our uses cases on top of it once they will be ready (lead by my colleague Daniele Pizzolli) and ii) start the implementation of the use-cases, lead by me.
15:39:24 <parus> Great!
15:39:40 <msimonin1> dsantoro: don't hesitate to ping us regarding EnOs if needed
15:40:00 <dsantoro> I will also update the slides to put clearer the network connection between edge and central cloud (as requested last time from adrien)
15:40:39 <dsantoro> sure, we are completing the setup of hardware, probably starting from next week Daniele Pizzolli will ask help (if needed). thanks a lot
15:40:50 <parus> #action dsantoro to update slides
15:41:15 <parus> #action dsantoro to explore ENOS and reach out to msimonin1 if needed
15:41:59 <parus> dsantoro: Beyond creation of VM are you going to be looking at other lifecyle functions
15:42:31 <parus> e.g.: maybe a destructive use case?
15:42:38 <dsantoro> @parus: can you please clarify ?
15:43:06 <parus> Well in your deck you did mention link failure as one important aspect.
15:43:34 <dsantoro> Yes of course in the 3rd UC we would like to isolate the edge node simulating a link failure between central cloud and edge node
15:44:02 <parus> OK.
15:44:45 <parus> Thanks. Regarding my deck. Jamecc and I have been in contact to try to validate some of the assumptions.
15:45:05 <parus> We have had one offline discussion and we are planning more.
15:45:07 <dsantoro> the application running on the edge node will save data locally, then later, when the connection will be up again data will be transitted back to the persitent storage in the cloud node
15:45:35 <parus> One key question is also regarding backhaul links and their capacity.
15:45:53 <parus> jamecc: do you want to comment further?
15:45:55 <jamemcc> I will put some time into trying to incorprate more of the 5G and Telco Fog Edge needs taht came out of the Boston presenation into our deck.
15:46:37 <parus> thank you. Should I capture that as an action?
15:46:39 <jamemcc> I also agreed to try to blow out some of the Control Plan Use Cases on slide 15
15:46:42 <jamemcc> Yes
15:47:08 <parus> #action jamecc to incorporate 5G and Telco Fog Edge needs Use cases
15:47:38 <parus> #action jamecc to work on Control plane use cases
15:47:55 <parus> Thank you!
15:48:01 <parus> Any other comments or inputs?
15:48:39 <parus> #topic open Discussion
15:49:14 <parus> we have a few minutes left. Is there anything else?
15:49:53 <parus> I propose we adjourn early then.
15:50:15 <msimonin1> thanks parus for chairing
15:50:21 <parus> Thank you all
15:50:22 <jamemcc> Thank you Paul-Andre - good leading
15:50:33 <parus> #endmeeting