21:00:31 <oneswig> #startmeeting scientific-wg 21:00:31 <openstack> Meeting started Tue Jul 25 21:00:31 2017 UTC and is due to finish in 60 minutes. The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:34 <openstack> The meeting name has been set to 'scientific_wg' 21:00:35 <oneswig> oh hi 21:00:47 <oneswig> #link agenda for today is https://wiki.openstack.org/wiki/Scientific_working_group#IRC_Meeting_July_25th_2017 21:01:28 <oneswig> Not a massive amount planned for discussion this week (phew) 21:01:29 <mrhillsman> o/ 21:01:38 <oneswig> Hi mrhillsman, good to see you 21:01:45 <trandles> hello 21:01:50 <oneswig> Hey Tim 21:02:29 <oneswig> martial: you online? 21:02:39 <martial> Hi Stig, yes just joined 21:02:45 <oneswig> The very chap :-) 21:02:47 <armstrong> Hello 21:02:51 <oneswig> #chair martial 21:02:52 <openstack> Current chairs: martial oneswig 21:02:55 <oneswig> Hi armstrong 21:03:15 <oneswig> OK lets get this show on the road 21:03:33 <oneswig> #topic SuperUser article review 21:03:54 <oneswig> OK, we had the interview, probably a couple of months back now 21:04:08 <oneswig> The copy is up for review 21:04:49 <oneswig> Were I a prepared man, I'd have it to hand.. 21:05:14 <b1airo> Hi 21:05:26 <oneswig> Hi b1airo 21:05:30 <oneswig> #chair b1airo 21:05:31 <openstack> Current chairs: b1airo martial oneswig 21:05:36 <trandles> https://docs.google.com/document/d/1k-lkP4mKNw9JxyDnCCxUmhjJm433zch5A4udmnJzRCw/edit 21:06:09 <martial> thanks Tim, I was looking it up :) 21:06:15 <trandles> np 21:06:31 <oneswig> Good work Tim, beat me to it 21:06:41 <armstrong> Can we add suggestion/contributions on the google-doc? 21:06:56 <martial> Yes, in the form of "suggestions" if you can 21:07:03 <armstrong> ok 21:07:38 <martial> “suggest edits” 21:11:32 <martial> (I have to say, I like the current version a lot) 21:12:47 <b1airo> We have a good looking University & Research track line-up for Sydney 21:13:35 <martial> blair: that sounds great 21:14:35 <oneswig> b1airo: good to hear it. How many slots to fill? 21:15:08 <trandles> don't think we know the slot # yet 21:15:34 <jmlowe> sorry I'm late, what did I sign up for in absentia? 21:15:37 <b1airo> No, I think that decision come later 21:16:08 <oneswig> jmlowe: just looking over this: https://docs.google.com/document/d/1k-lkP4mKNw9JxyDnCCxUmhjJm433zch5A4udmnJzRCw/edit 21:16:50 <martial> jim: you are judging the HPC hackaton live at the Sydney summit (it is a 72 hour events, caffeine to be provided) ... thanks for "volunteering" 21:17:30 <b1airo> Lol 21:17:30 <jmlowe> NAVY - never again volunteer yourself 21:18:55 <martial> jim: kidding of course :) 21:19:42 <oneswig> martial: were you also kidding when I got signed up? :-) 21:21:03 <martial> stig: depends, are you asking about the ORC meeting in Amsterdam? 21:21:31 <martial> stig: the other one is totally different, it is not an *HPC* hackaton :) 21:21:51 <oneswig> Ah, so true 21:21:54 <martial> stig: coffee to be provided (hopefully) as well 21:22:14 <jmlowe> I'd love to go to Amsterdam, I'm not going to that meeting, as I told a colleague I wouldn't go that meeting if it was downstairs 21:22:29 <oneswig> I might need something stronger! 21:22:59 <oneswig> martial: Not Amsterdam, was the hackathon Flanders mailed about 21:23:26 <martial> stig: yes, I know, and yes, you are the judge :) 21:23:51 <martial> everybody happy with the state of the "interview" document? 21:24:03 <b1airo> I can't keep up with this conversation! 21:24:30 <jmlowe> I'm relatively happy with the interview 21:24:38 <trandles> +1 21:24:41 <b1airo> I had first review of it a week ago, so assume you've undone all my edits now :-) 21:24:42 <jpr1> q: should the intro sentence have one line about what openstack (possibly as part of the first sentence) 21:25:18 <b1airo> jpr1: it will be posted on the OpenStack site so maybe redundant? 21:25:22 <martial> jpr1: it is for publication in the SuperUser magazine, so I am not sure 21:25:23 <trandles> jpr1: it's for Superuser so the audience _should_ be familiar 21:25:29 <jpr1> and possibly put "why openstack for research" ahead of "why care about ... working group" 21:25:39 <jpr1> ok. that audience likely has a clue. ;) 21:26:05 <jpr1> pardon being out of context at the moment 21:26:11 <b1airo> jpr1: yes I agree, I think it could talk about both why and who 21:26:26 <martial> jpr1: no worries, but thank you for the suggestion, it is very relevant 21:27:02 <oneswig> OK - publish and be damned? Last changes please. 21:27:06 <jpr1> just reading this from point of view of someone who might see just this doc, ie. a url in an email 21:27:37 <jpr1> no worries on any of the above late sideline commentary btw 21:28:07 <jpr1> looks really good btw 21:28:09 <martial> stig: Gene is going to love the extra suggestions :) 21:28:39 <martial> stig: with Flanders being ready to push to the editor in chief 21:28:54 <martial> stig: we ought to warn them both 21:29:22 <oneswig> martial: OK, I'll mail them now 21:30:13 <martial> stig: just did so 21:30:35 <oneswig> ah, so you did :-) 21:30:59 <oneswig> OK - move on? 21:31:12 <martial> +1 21:31:29 <oneswig> #topic SC2017 BoF 21:31:53 <oneswig> OK, previously we put together a draft BoF proposal 21:32:04 <oneswig> Based on SC2016 21:33:02 <jmlowe> an all new one for SC2017? 21:33:19 <oneswig> I thought so but can only find the SC2016 one - b1airo? 21:33:28 <oneswig> was it an etherpad? 21:33:53 <trandles> I don't recall one for SC17 21:34:11 <b1airo> Hmm yes I think we may have minted an etherpad for brain storming 21:34:24 <oneswig> #link I found this https://etherpad.openstack.org/p/SWG-SC17 21:34:33 <b1airo> But we should just work from the sc16 one 21:35:07 <b1airo> I'm on my phone so bit handicapped sorry 21:36:09 <oneswig> #link updated for SC17 https://docs.google.com/document/d/1wMk1qoytJuIY7REzVc_eRpkw4w1qC_-hQf1xnvwb_b8/edit?usp=sharing 21:36:20 <jmlowe> I put in my request to the IU boot committee for a set of talks 21:38:37 <trandles> Fun fact: My Sydney travel was approved quickly but SC17 in Denver is up in the air. Who'd have thunk it...I might not make SC... 21:39:20 <b1airo> trandles: odd, are people going off it in favour of ISC? 21:40:06 <jmlowe> nobody from the US goes to ISC 21:40:09 <trandles> No, it's that there are usually a huge number of requests by the DoE labs so HQ doles out a set number/lab 21:40:37 <jmlowe> I'm checking with Evan Bollig as a sub for Jonathan Mills on the BoF 21:40:42 <b1airo> Maybe we could make the BoF title: Models for HPC with OpenStack ? 21:40:50 <trandles> once LANL has their number, my division will decide who gets to go 21:41:02 <b1airo> And aim to get a balanced discussion across different architectures 21:41:03 <jmlowe> b1airo: +1 21:41:55 <trandles> does "Models" imply "use cases?" 21:43:40 <jmlowe> I wouldn't say that, you can do the same use case across different models 21:46:18 <b1airo> Would be nice to also manage to have some focus on apps/users 21:48:56 <trandles> IMO users are more effective advocates than admins for emerging tech and HPC 21:49:54 <b1airo> It's somewhat difficult in OpenStack land as it's so ops & architecture heavy 21:50:09 <bollig> sorry jmlowe: I’m in. either I or another from MSI will present 21:50:29 <trandles> admins have a reputation for building neat toys that don't deliver a clear win for users 21:50:31 <jmlowe> ok, I'll pencil you in for the draft 21:51:15 <martial> I would love to say I can also make it, but I am in still unsure if I will be able to be present at either event 21:51:27 <trandles> If we can get one or two users with clear stories about how cloud delivered real science it would drive home the point 21:51:41 <b1airo> jmlowe: I could do an update on our talk too, but I'm happy to be relegated if there is more variety available 21:52:05 <jmlowe> Maybe I can recycle materials from our first annual nsf review, it was half a day of users detailing how they used jetstream 21:52:33 <bollig> we’ve got users focused on protected NIH data (dbGaP) and genomics workflows 21:53:16 <b1airo> Protected = only shared with Putin? :-P 21:53:25 <jmlowe> that's some really interesting stuff, anything with sensitive data is always going to get attention 21:53:37 <b1airo> (sorry, couldn't help it) 21:53:43 <jmlowe> too soon 21:53:56 <bollig> ha! 21:54:34 <b1airo> But yes, we have these use-cases too - Monash is working to launch a Health Data Platform 21:54:39 <trandles> bollig: we had a similar use case with NIH and cancer genomics data 21:54:50 <martial> stig, blair, any issue if I add my name on the sponsor list in case I can get it approved? 21:55:02 <b1airo> Which will include cloud based safe havens for analysis workloads 21:55:02 <oneswig> of course - go ahead 21:55:20 <oneswig> b1airo: might be interesting to discuss offline 21:55:53 <trandles> public cloud-based dataset access is one of the strongest desires our users have 21:56:02 <b1airo> Sounds good oneswig 21:56:32 <b1airo> By the way, anyone coming to Sydney should think about visiting Melbourne the week prior 21:56:37 <martial> stig: done, thanks 21:56:48 <b1airo> We will happily host something at Monash 21:57:00 <martial> b1airo: I am hoping to be able to do that 21:57:16 <trandles> b1airo: unfortunately I can't make it the week before 21:57:24 <oneswig> b1airo: you going to stick some prawns on the barbie? 21:58:02 <b1airo> Maybe a few snags oneswig 21:58:24 <b1airo> Wow 2 minutes to the hour 21:58:35 <b1airo> #topic AOB 21:58:38 <jmlowe> that went quick 21:59:46 <b1airo> I have a code that takes 3x as long without hugepages in the guest as with 21:59:51 <oneswig> No other business for me - been crazily busy this week 22:00:12 <b1airo> Now trying to learn how to get useful data from perf-kvm 22:00:14 <oneswig> b1airo: TLB thrashing? 22:00:33 <oneswig> b1airo: will repay you in the end, no doubt 22:00:37 <jmlowe> I've got one for you, virtual routers change the source from outside packets to be from itself 22:00:37 <b1airo> Must be I think, but I'm surprised at the magnitude 22:01:11 <trandles> jmlowe: that sounds like a NAT not a router... 22:01:22 <oneswig> OK better wrap up y'all 22:01:25 <jmlowe> can't figure out how he made it do that 22:01:31 <b1airo> Really?? How long have we been doing NAT now - want that sort of problem solved years ago 22:01:42 <jmlowe> continue in regular channel 22:01:54 <oneswig> Thanks all 22:02:00 <oneswig> #endmeeting