11:00:54 #startmeeting scientific-wg 11:00:55 Meeting started Wed Sep 27 11:00:54 2017 UTC and is due to finish in 60 minutes. The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot. 11:00:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 11:00:58 The meeting name has been set to 'scientific_wg' 11:01:02 ahoy there 11:01:15 o/ 11:01:20 #link Agenda for today https://wiki.openstack.org/wiki/Scientific_working_group#IRC_Meeting_September_27th_2017 11:01:28 Hi daveholland, thought you were on holiday? 11:01:42 btw what is your job title, for attribution purposes? 11:01:45 I wish ;) 11:01:55 good morning 11:02:00 Hi martial_ 11:02:02 #chair martial_ 11:02:03 Current chairs: martial_ oneswig 11:02:04 "Principal Systems Administrator" (I think that's an anagram of "dogsbody") 11:02:04 allo 11:02:17 OK, thanks, noted 11:02:22 #chair b1airo 11:02:23 Current chairs: b1airo martial_ oneswig 11:02:32 morning Blair 11:02:34 o/ 11:02:38 hi b1airo, morning 11:02:42 Hi johnthetubaguy :-) 11:02:52 * johnthetubaguy waves back at oneswig 11:02:55 evening :-) 11:03:16 b1airo: we will get to play in your timezone sandbox soon hopefully :) 11:03:30 OK - on which subject.... 11:03:34 oneswig, johnthetubaguy - are you guys waving at each other from opposite desks now? 11:03:39 #topic Sydney hackathon 11:03:54 b1airo: only virtually :-) 11:04:10 oneswig: what about the hackaton? :) 11:04:11 yeah, multiple chat tools maybe 11:04:14 Saw him yesterday at OpenStack UK though! 11:04:56 Ah, yes. The hackathon. Is looking for mentors and supporters. Its the weekend before the summit, starting Friday 3pm (in Sydney) 11:04:58 ICQ 4eva! 11:05:11 #link http://hackathon.openstack.org.au/mentors/ 11:05:31 Also seems to mention "Mentors will also receive a free pass to the Sydney OpenStack Summit" 11:05:37 just leaving that htere 11:05:54 oneswig: yes I confirm 11:06:27 martial_: are you a mentor? 11:06:33 heh, i'll ping a couple of people that might be good candidates if they are in the country at the time 11:06:55 unfortunately it won't work for me, can't get to Sydney until Sunday evening 11:07:19 You and Wojciech too! :-) 11:07:31 oneswig: I registered to be one, and David put me in contact with the lead mentor 11:07:57 oneswig: are you still judging? 11:08:29 I have not heard confirmation on that but am assuming I'll be involved somewhere 11:08:52 Should be fun :-) 11:09:14 sounds fun to me, indeed :) 11:09:25 OK, move on? 11:09:36 #topic Open Research Cloud 11:09:58 next ORC meeting starts about now in Amsterdam. 11:10:00 well they are still setting up, I am on the GoToMeeting : 11:10:02 ) 11:10:10 Ahah, multi-tasking! 11:10:32 well I can listen onto one and type in this one :) 11:10:35 wow, you can live-IRC it for us 11:10:49 It's almost like we are in the room :-) 11:10:55 should start in 20 minutes 11:11:05 the video feed is already up ... 11:11:32 Tomorrow there is a technical session, in which there is a section designated "Scientific WG perspectives" 11:11:34 oneswig, martial_ - are you across any of the plans/intentions/pie-in-the-sky ideas for ORC meeting in Sydney ? 11:11:52 I saw some meeting invites that's all 11:11:59 oh really? do we need to come up with a perspective then!? 11:12:21 Well this is it, I've got writers block 11:12:38 Martial are you covering the NIST project? 11:13:00 b1airo: it will be a panel I believe more than anything else 11:13:21 oneswig: the IEEE / NIST P2302? 11:13:32 the only perspective we have to offer is "please don't invent another standard" (for auth etc) 11:13:33 martial_: are there others? 11:13:55 b1airo: if there are slides to be had, it might not be crazy :) 11:14:05 daveholland: seconded... 11:14:06 oneswig: now that I think, likely no :) 11:14:19 oneswig: Craig Lee from IEEE will be there so they are covered 11:14:29 I will cover the P2302 for the foundation for now 11:14:50 (per my conversation with David) 11:15:37 OK. We are joining remotely for that session and I was thinking of summarising the new book chapter, however it's written by Enol who's hosting this session so he's probably already got it well covered... 11:16:04 Speaking of the book 11:16:22 Hi hogepodge - welcome and good morning! 11:16:23 Ildiko Vancsa from the Foundation is at the ORC, so she is covering for OpenStack 11:16:34 It's due this Friday no exceptions all content and finished diagrams 11:16:49 hogepodge: good morning, and yes we confirm 11:17:07 hogepodge: what timezone on Friday? :) 11:17:14 it is. I believe I only have the new federation chapter to review 11:17:44 Let's say noon central time for Kathy's sake 11:18:07 hogepodge: works for me. Did you see, we gathered some proposals for cover art? 11:18:20 oneswig: Mike is finishing his "Research in Production" chapter and I am reviewing it to add 11:18:22 #link down here somewhere https://etherpad.openstack.org/p/scientific-computing-book-update-2017 11:18:41 martial_: is that a chapter or a case study? 11:18:51 They need to verify that they license is good for the cover, I'll forward it to the team 11:19:27 hogepodge: the images have been offered on the terms previously described, I think they are clear on that. 11:19:33 Ok 11:19:47 oneswig: a little of both, it is a longer version of his lighting talk from Boston 11:20:48 on the cover art... 11:21:21 martial_: to fit as a chapter it'd want a broad range of case studies 11:21:54 i like the SKA idea, though i do wonder whether it might raise a few eyebrows to have pics of SKA antennae in Australia given that so far the Aus contingent seems to be traditionally HPC focused... 11:22:34 b1airo: a good observation 11:23:15 I gave them access to our OpenStack SKA system just last week however 11:23:27 oneswig: good observation, I am passing along this information 11:23:39 certainly the Pawsey Supercomputing centre still uses OpenStack, but at this point I understand it to be a side-show (i could be wrong/out-of-date here though) 11:23:43 they are now tainted with the corruption of software-defiend infrastructure 11:24:05 haha, good work 11:24:12 b1airo: specifically this was a team from UWA, might be connected, might not 11:25:27 The SKA Organisation contributed the images, I don't think there's any perceived issue from them 11:25:57 cool, no probs then 11:26:00 #topic Book 11:26:16 (totally forgot to change the topic :) ) 11:27:04 I think the main body of work is ready, the additions appear close and one schedule 11:27:29 hogepodge: would it help to have access to get started on the next stage for the completed sections? 11:30:08 Well I think we are on schedule, anyway, providing the sky doesn't fall in in the meantime... 11:30:21 move on? 11:30:42 #topic Sydney Forum 11:30:54 I must confess I have not been keeping up with this 11:31:03 me either 11:31:27 the close date for Forum topic submissions is 29th 11:31:52 unless I misunderstood, they request from us a list of the talks/sessions we intend to have? 11:32:14 Which ML is it discussed on? 11:33:06 #link https://etherpad.openstack.org/p/Scientific-Sydney17-Forum 11:33:10 martial_, not quite - the idea for the process is that the community (including WGs/SIGs) first brainstorm internally and that provides a way to shake out duplicates and form up ideas, then session leaders go ahead and put the proposals in 11:34:07 that brainstorming phase is more-or-less over already, so i think we need to figure out whether we have any good ideas for Forum sessions between now and tomorrow and then go ahead and then get someone to submit an abstract 11:34:22 oneswig: I got an email from the "User-committee" ML 11:34:53 thanks martial_ b1airo, now I recall the process. 11:35:05 our plan is to propose a BoF and lighting talk sessions so far 11:35:09 So what to advocate.. 11:35:12 the sessions are certainly picked by representatives from both the UC and TC based on the submitted ideas, so that sounds right 11:35:13 i would like to have another session on "opportunistic" instances, but i feel a bit phoney doing so given i never properly followed up on the same session from last Forum 11:35:42 honestly, no one seems to be working on much of that right now, I wrote up the spec from the discussions 11:35:56 similar issues with quotas right now 11:35:58 martial_, no we already have a BOF and lightening talk session in the main Summit's Working Group agenda 11:36:25 b1airo: so oneswig is not the only one confused I confess :) 11:36:26 b1airo: I think it's a missing piece. It would be surprising if the public cloud wg weren't already onto this too 11:37:14 @oneswig: re ORC "slides would be good" 11:37:31 OK thanks martial_, noted. 11:37:51 johnthetubaguy, yeah and i meant to try and continue that conversation with you on the draft spec 11:38:43 I think something useful is expressing the relative importance of a top few efforts, and agreeing that with other SIGs and WGs 11:38:45 johnthetubaguy: would it be worthwhile to add an item on Ironic resources and traits 11:38:50 lately i have been wondering if there might be a simple way to get 80% of what i want by just writing an external service that arbitrated access to some defined free capacity 11:39:34 oneswig: unsure who would be there, but possibly on cross project scheduling requirements in general, just making sure everyone is happy with the placement plan overall? 11:39:43 hello 11:39:47 sorry I am late, but I arrived 11:39:56 Hi zioproto! 11:40:01 Just talking forum topics 11:40:04 @stig: agenda has shifted to 13:15 to 14:15 for "Technical Perspectives from the SWGs" 11:40:06 o/ zioproto 11:40:20 (Amsterdam time) 11:40:22 b1airo: thats a lot of what blazar is doing, there was PTG discussion that ended with just increase the quota for folks who request the feature for now 11:40:27 OK thanks martial_ 11:40:46 guys for Sidney we are looking for moderators for the Forum 11:40:55 anyone wants to jump in ? 11:40:58 Moderators should propose the sessions they sign up for to http://forumtopics.openstack.org/ 11:41:14 johnthetubaguy, as in have Blazar increase the quota? 11:42:01 zioproto: good timing and welcome 11:42:07 b1airo: I think I just confused things, I mean something like blazar would do the increase in quota, use tags on instances to manage the reaping of instances, etc (i.e. build a prototype outside Nova for now, quota is the main sticking point) 11:43:03 so I am still confused about what you need from us related to our forum proposals. I apologize in advance, but clarification on the purpose and example of topics would be appreciated 11:43:11 ^ zioproto 11:43:39 johnthetubaguy, yep thanks, think i have a vague picture of what you're talking about - would have to be quota plus aggregates for reservation of some kind though, for us quota is fairly useless as we have *way* more quota allocated than we have actual capacity 11:44:16 b1airo: OK, so yours must be a different use case to the one in the spec, needs more follow up another time I guess 11:44:23 martial_: I am looking for moderators for the sessions that appear at http://forumtopics.openstack.org/. After you login with your openstack account you should be able to see the session names 11:44:28 martial_, you could look at the Boston Forum agenda 11:45:30 for me the forum is all about getting Users and Developers together agreeing the priorities and requirements for the next cycle, keeping in sync with what has happened so far, etc 11:45:50 #link https://wiki.openstack.org/wiki/Forum#At_the_Forum 11:45:58 the sessions that connection on going efforts with interested folks, worked really well in Boston 11:46:01 Do we know the scheduling of forum sessions yet? 11:46:13 #link https://etherpad.openstack.org/p/SYD-UC-brainstorming 11:46:14 b1airo: https://www.openstack.org/summit/boston-2017/summit-schedule/global-search?t=The%2BForum 11:46:27 some of the relative priority setting stuff seemed to work quite well in Boston too 11:46:54 like ignore all these things, this is the one thing that is blocking everything (application credentials came out of that sort of discussion) 11:47:04 not sure if that helps with context 11:47:34 thanks johnthetubaguy I think it does. 11:47:38 actually looking at the list of topics already suggested and hearing johnthetubaguy mention getting Ops & Users & Devs together reminded me that we no longer seem to have things like the Ops sessions of old 11:48:10 there is more context here: https://ttx.re/splitting-out-design-summit.html 11:48:10 maybe an HPC Ops then? 11:48:19 the idea is to work together to come up with join sessions 11:48:27 rather than have ops come to dev sessions, and v.v. 11:48:41 joint^ 11:48:43 johnthetubaguy: seen this one? http://forumtopics.openstack.org/cfp/details/12 11:49:20 yeah i understand the motivation to jointly get the agenda together, but in the process we seem to have lost what were some very well attended sessions previously, e.g., "Ceph Operators" 11:49:43 oneswig: I haven't reviewed any of them honestly, mostly as I am not choosing and not attending 11:49:44 seems important to have a home for those things too 11:49:56 oneswig: that sounds like pre-ptg speak 11:50:16 oneswig: I guess its more a follow on 11:50:43 johnthetubaguy: right, it does read like that. 11:51:02 b1airo: Is there not a Ceph BoF? Really ought to be, I agree 11:51:21 yeah i will propose and see what happens 11:51:25 so looping back, I like that session because it is getting feedback around existing ongoing work, which is good 11:51:41 Would love to share our current weirdness with the room :-) 11:52:03 so "HPC Ops Unite" & "Ceph BoF" ? 11:52:32 Isn't HPC Ops Unite the tagline for our SIG? 11:53:00 oneswig: yes but we do not have a WG meeting anymore, just BoF and Lighting Talk 11:53:24 johnthetubaguy, one observation from Ops land is that it seems like some of those sessions can be hamstrung because there are so few operators with experience of current releases - so they haven't even hit the beginning of the features you want feedback on 11:53:30 so using the forum to give people a place to meet the community is still a good opportunity I think 11:53:55 b1airo: very true 11:54:00 ie shifting from a "planned work" toward a "come and join us" 11:54:14 well volunteered martial_ :-) 11:54:48 oneswig: I will take the "HPC Ops Unite" if somebody else take the "Ceph BoF" ? 11:55:00 might be a point worth throwing into the current "Simplification" thread on os-dev, and certainly makes an argument for closing the gap toward less dev reliance on devstack 11:55:19 I am already speaking every day of the summit and will blow a gasket if I take on anything more 11:55:42 martial_, yep just checking the agenda to ensure there is not already such a Ceph session (i know there is a CephFS panel, because we are on it...) 11:56:46 oneswig: added it to b1airo https://etherpad.openstack.org/p/Scientific-Sydney17-Forum 11:56:53 b1airo: thanks as well 11:56:56 OK, thanks martial_ b1airo - anything more to add on proposals? 11:57:05 bloody google, searching for "openstack sydney ceph bof", all i get is spinics links to ML posts of me asking about this a couple of months ago 11:57:19 a feedback from a dev side of things, it was really hard with lots of WG meeting in Boston 11:57:24 In a silent room, echoes... 11:57:27 I had no idea which I should attend 11:57:43 johnthetubaguy: what do you mean exactly? 11:57:58 the ones I did often spoke about the things that were mentioned in lots of the other rooms, I can't remember which topic that was 11:58:04 oh wait that is not a topic reques :) 11:58:06 +t 11:58:12 so lets be specific on quotas 11:58:19 maybe we could suggest a "top pain points for scientific openstack" session ? 11:58:23 lots of folks spoke about quota problems in different rooms 11:58:34 b1airo +1 11:58:48 johnthetubaguy, yep good idea 11:58:50 the problem as dev is I didn't know what room to go to talk about quotas, beyond the quota session 11:59:02 so we fragment the discussions 11:59:14 I think it would be better if the sessions broke out of all our silos 11:59:16 get out ahead of the game on that one and make sure there is something on the agenda to point people at where that will (at least) all get collected on a single etherpad 11:59:31 sounds good b1airo 11:59:32 now that is crazy hard... its just the sorts of sessions that seemed to most successful in Boston 12:00:34 the successful ones also really need to have at least draft blueprints that attendees can study up on prior 12:00:34 OK we are out of time 12:00:41 to be clear the etherpads doesn't really too much here, you still get three fragmented discussions instead of getting everyone interested in the same room 12:00:49 oops, we are over 12:01:07 I had one piece of AOB to mention 12:01:26 no one is waiting for the room oneswig so may as well go ahead :-) 12:01:37 please keep in mind if you want to moderate a session to add yourself up into the web tool ! 12:01:40 priteau and his wife had a healthy baby girl last week, all going well 12:02:27 great news 12:02:32 congrats ! 12:02:33 I like ending on good news, congrats to priteau 12:02:43 I shall pass that on! 12:02:57 good news unless you like sleeping ;-) 12:02:57 Could continue the forum disucussion in #scientific-wg 12:03:19 oneswig +1 12:03:20 yep i will hang around a while - got cable lengths to figure out... 12:03:30 ooh, shopping... 12:03:41 ok, lets clear the channel here 12:03:44 #endmeeting