15:00:26 <ttx> #startmeeting large_scale_sig
15:00:26 <opendevmeet> Meeting started Wed Aug  4 15:00:26 2021 UTC and is due to finish in 60 minutes.  The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:26 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:26 <opendevmeet> The meeting name has been set to 'large_scale_sig'
15:00:32 <ttx> #topic Rollcall
15:00:34 <oneswig_> hi
15:00:42 <ttx> oneswig_: hi!
15:00:47 <ttx> Who else is here for the Large Scale SIG meeting?
15:00:51 <belmoreira> o/
15:01:01 <ttx> Hey Belmiro, not vacationing yet?
15:01:16 <ttx> that said, with the weather we have...
15:01:22 <belmoreira> looking forward to the end of the week :)
15:01:35 <ttx> belmoreira: where are you going?
15:02:17 <ttx> pinging amorin
15:02:33 <ttx> Our agenda for today is at:
15:02:34 <belmoreira> I will go to Portugal and try to have some weeks next to the sea
15:02:34 <ttx> #link https://etherpad.openstack.org/p/large-scale-sig-meeting
15:02:57 <ttx> nice, going to Southwest France / Northeast Spain myself
15:03:16 <ttx> Alright let's get going
15:03:17 <belmoreira> great
15:03:20 <ttx> #topic Next OpenInfraLive session: Discussing software-defined supercomputers (August 26)
15:03:32 <ttx> For this one, given that our usual crew is not available, I opted for something slightly different
15:03:37 <ttx> A whole separate facet of "large scale": software-defined supercomputers! :)
15:03:49 <ttx> oneswig_ volunteered to lead a discussion on this and opened his address book
15:03:57 <ttx> So far we secured 4 guests:
15:04:03 <ttx> Sadaf Alam (CTO of CSCS, the Swiss national supercomputer centre)
15:04:11 <ttx> Happy Sithole (Director of Center for High Performance Computing (CHPC) South Africa)
15:04:19 <ttx> Steve Quenette (Deputy Director, Monash e-Research Centre)
15:04:27 <ttx> and of course Stig who will host the discussion.
15:04:40 <belmoreira> this will be a great session!
15:04:41 <oneswig_> ttx: do you have a good example of topics for a lively discussion from previous sessions?  Anything you'd like to see included?
15:04:47 <belmoreira> thanks onewig_
15:04:48 <oneswig_> hope so belmoreira!
15:05:00 <ttx> we can brainstorm that a bit, let's talk speakers first
15:05:02 <ttx> We have two wildcards: Gregory Orange from Pawsey, and Dorah Thobye (also from CHPC)
15:05:12 <ttx> And finally we also have two as-yet-unanswered invitations: Jonathan Mills (NASA Goddard) and Paul Calleja (Cambridge)
15:05:22 <ttx> I'd like to give answers to Greg and Dorah today if possible.
15:05:34 <ttx> But I would also like to give Jon and Paul a chance to participate, while keeping the number of speakers at 6 max
15:06:03 <ttx> So maybe the simplest is to reply "no" to our wildcards, keep the door open for Jon and Paul. Worst case scenario if they can't make it, we'll have only 4 guests.
15:06:09 <ttx> Stig: does that make sense?
15:06:20 <oneswig_> I think Dorah's a substitute for Happy, if he can't make it.  I don't think you would need both involved.
15:06:27 <ttx> cool
15:06:34 <oneswig_> But yes that makes sense
15:07:03 <oneswig_> Greg is very kind to offer to stay up past midnight.
15:07:22 <ttx> I'll be around to do the intro and outro, as well as selecting pertinent questions during the show. That way you can fully focus on leading the discussion.
15:07:45 <oneswig_> That would be a great help
15:07:47 <ttx> regarding the discussion script, the idea is to work on it between now and the event... And then we can present it to the guests just before the show.
15:08:03 <ttx> belmoreira: any tips?
15:08:23 <ttx> I think the general idea is to have a series of base questions, but a lot of flexibility based on how the discussion goes
15:08:44 <ttx> Start with intros and give guests the possibility to explain what they do
15:09:14 <oneswig_> I hope we will run out of time, rather than run out of content.  Usually it works that way.
15:09:21 <ttx> then very open questions like "unique challenges of that use case", "what role openstack plays" and then dive deeper
15:09:21 <oneswig_> Does the format include questions from the audience?
15:09:51 <ttx> yes, we take questions from the audience. usually we try to take at least one early in the show to show that people can actually ask questions and be featured
15:10:01 <ttx> You can review the last one we made as an example:
15:10:21 <ttx> https://www.youtube.com/watch?v=G7oN2XdI__k
15:10:27 <oneswig_> Super, thanks.
15:10:51 <ttx> The platform we use allows to collect questions from all the streaming places (LinkedIn, YoutUbe...) and see them all
15:11:07 <ttx> then you can show the question on the screen while answering it
15:11:13 <belmoreira> having ttx monitoring the  live questions is great... maybe you should also join with the questions during the discussion if you think they are pertinent... in my case it was very difficult to follow the discussion and chat... we should avoid leaving all the live questions to the end
15:11:41 <ttx> like https://youtu.be/G7oN2XdI__k?t=1572
15:12:23 <ttx> Yeah the trickiest part is getting the discussion lead to acknowledge and introduce a selected question
15:12:38 <ttx> We use the tool's internal chat for that
15:13:39 <oneswig_> And the moderator puts the question up on the screen?
15:13:46 <belmoreira> ttx what you just mention earlier it great... if you see a good question just stick it in the screen and is addressed next
15:13:46 <ttx> but at least it's a lot simpler than when a single person monitored questions, selected them and followed the discussion at the same time
15:14:19 <ttx> our producer folks will put it on screen when we tell them
15:14:50 <oneswig_> That sounds easier, one less thing to worry about!
15:15:03 <ttx> In terms of prep you need: a loose script to follow in case the discussion does not naturally develop, a list of messages banners you'd like to show.
15:15:07 <mdelavergne> (hi, sorry for being late!)
15:15:38 <ttx> It's a lot easier than it sounds really. Selecting the right people to have the discussion is actually the hardest part :)
15:15:57 <ttx> If you pick the right people, the discussion naturally happens and makes the best shows
15:16:13 <ttx> the script is there more as a backup in case the discussion stalls
15:16:19 <oneswig_> I think we've got a fairly chatty bunch with no shortage of opinions between them :-)
15:16:23 <ttx> or in case there is a question you absolutely want to tackle
15:17:15 <ttx> That's an example of a prepared banner that we can display: https://youtu.be/G7oN2XdI__k?t=2726
15:17:37 <ttx> Those it's good to have a list in advance so that we can enter them ahead of time
15:17:47 <ttx> but we can also make them on the spot
15:17:49 <oneswig_> Is there an etherpad already for script ideas, or should I start one?
15:18:15 <ttx> oneswig_: in Ashlee's message you have a link to a collab doc on Google docs
15:18:31 <oneswig_> ah, great. Thanks ttx
15:19:20 <ttx> You can dump your thoughts there, and we can finalize them the week of the event
15:19:47 <ttx> Any other question on how it all works?
15:20:07 <oneswig_> I think that's plenty for now thank you
15:20:54 <ttx> We also have the possibility to show a few slides at the start to introduce the topic, but it's also OK if we dont have any
15:21:12 <oneswig_> Good to know
15:21:23 <ttx> In our case it feels like the real highlight will be the exciting science those people do in connection with openstack, so I'd rather talk about that instead
15:21:57 <ttx> unless you think we should ask each of them a single slide to present their setup, or something
15:22:30 <ttx> (in which case we'd have to ask them for it)
15:23:09 <ttx> I'm fine with whatever, especially since I won;t be much around between today and the show to help gathering them
15:24:01 <ttx> ok, maybe we can move on to next topic?
15:24:08 <oneswig_> I am not sure.  I don't want to impose, and it might not add much.  There may be options there.
15:24:17 <oneswig_> go ahead
15:24:26 <ttx> #topic Progress/Blockers
15:24:52 <ttx> With most of the crew out of office, I suspect we don;t have much Large Scale SIG progress to report.... anything belmoreira ?
15:25:18 <belmoreira> sorry, will try to have it ready next time
15:25:34 <ttx> OK let's jump directly to...
15:25:37 <ttx> #topic Next meeting
15:25:47 <ttx> Our next scheduled IRC meeting will be... August 18. Should we have one then?
15:25:53 <ttx> That's the week before the OpenInfra Live, so could make a good last-minute org meeting
15:26:38 <ttx> I'm also very happy to skip it if nobody's around :)
15:26:53 <belmoreira> I will only be around in September
15:27:14 <ttx> oneswig_: do you want to coordinate then, or shoudl we just do emails?
15:27:26 <oneswig_> I should be around if there are last-minute issues
15:27:49 <ttx> How about we do emails, and jump on IRC/Zoom if necessary
15:28:00 <oneswig_> Good plan
15:28:07 <ttx> Alright, Aug 18 is a skip then!
15:28:11 <ttx> #info next IRC meeting on Sept 1st
15:28:29 <ttx> Our next open "Large scale openstack" slot for OpenInfra Live will be Oct 21, so that gives us time to organize.
15:28:39 <ttx> We could do the Inception episode or the Monitoring episode, depending on who can make it.
15:29:08 <ttx> belmoreira: you can mark your calendar already :)
15:29:24 <oneswig_> belmoreira: we've experimented with nova cell controllers in VMs
15:30:06 <oneswig_> but to avoid the recursion, they all reside in one cell that has a physical cell contoller.
15:30:13 <ttx> #info Next OpenInfraLive slot (after Aug 26) will be Oct 21
15:30:33 <ttx> Alright... Anything else, anyone?
15:30:40 <belmoreira> oneswig_ ok
15:31:09 <ttx> belmoreira just embraces the recursion
15:31:33 <oneswig_> ... but does the recursion embrace back
15:31:57 <ttx> that's one great mystery of the universe. Now that the higgs Boson is not
15:32:24 <belmoreira> it's even better when running OpenStack in k8s clusters deployed by magnum :)
15:32:38 <ttx> my head hurts
15:32:56 <ttx> Alright, Boson jokes means the meeting is probably over
15:33:01 <ttx> Thanks everyone
15:33:06 <oneswig_> thanks!
15:33:12 <belmoreira> thanks
15:33:21 <ttx> #endmeeting