21:00:25 <oneswig> #startmeeting scientific-sig
21:00:25 <opendevmeet> Meeting started Tue Oct 12 21:00:25 2021 UTC and is due to finish in 60 minutes.  The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:25 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:25 <opendevmeet> The meeting name has been set to 'scientific_sig'
21:00:38 <oneswig> aohy
21:00:44 <oneswig> I mean, ahoy...
21:01:09 <martial> hello as well :)
21:02:19 <oneswig> Hey martial
21:02:21 <oneswig> #chair martial
21:02:21 <opendevmeet> Current chairs: martial oneswig
21:02:50 <oneswig> I've been enjoying a hands-on week this week, how about you
21:03:31 <martial> still trying to figure out SC21
21:03:47 <martial> before I forget the Etherpad link we discussed is at
21:03:56 <martial> #link https://etherpad.opendev.org/p/202110-PTG_ScientificSIG_LightingTalks
21:04:13 <oneswig> Nice, thanks martial
21:04:20 <martial> it is very much a to fill blurb but it is ready
21:05:06 <oneswig> Have you thought of a talk topic yet?
21:06:56 <martial> for the lighting talks? not yet
21:06:58 <oneswig> I've recently done some work on storage performance benchmarking which might be worth sharing
21:07:18 <martial> that sounds really cool
21:07:28 <martial> I just posted the link on our slack as well
21:07:41 <b1airo> hey gang. lurking - in another meeting as usual (and catching up on leave)
21:08:16 <oneswig> Hi b1airo, morning
21:08:21 <oneswig> #chair b1airo
21:08:21 <opendevmeet> Current chairs: b1airo martial oneswig
21:09:09 <martial> welcome Blair, any topic you are presenting ?
21:09:44 <oneswig> b1airo: we are talking lightning talks
21:09:58 <trandles> hi folks...also lurking
21:10:09 <oneswig> hi trandles (same question to you...)
21:10:48 <b1airo> i could maybe volunteer Sean for a lightning talk
21:11:24 <b1airo> (overview of our new OpenStack based "FlexiHPC" platform)
21:11:39 <trandles> Unfortunately not. I could drone on for an hour about what happens with a program manager quits and the new one comes in and halts everything because they have a different vision than the last person.
21:12:04 <trandles> s/with/when/
21:12:53 <oneswig> trandles: sympathies
21:13:20 <oneswig> What about when you have two program managers?  It's like having two watches
21:14:23 <trandles> Oh, I haven't experienced two program managers on the _same_ program. That doesn't sound like fun.
21:14:52 <oneswig> It's generally true that you can have too much of a good thing
21:15:53 <oneswig> I don't think that's a topic for a lightning talk however
21:17:49 <oneswig> This week, Michal from our team is joining OpenInfra live to talk about scaling Neutron
21:18:01 <b1airo> :'-D
21:18:44 <oneswig> It should be a good session
21:23:54 <oneswig> I'm working through the OpenStack runes for using erasure coded pools with RBD (Glance, Cinder-Backup in this instance).  Surprisingly lightly documented...
21:24:51 <martial> the code it is strong with this one :)
21:28:13 <b1airo> is there anything particularly special in that regard oneswig ?
21:30:20 <oneswig> Just unfamiliar with the magic needed to have an EC data pool and a replicated meta pool.
21:32:23 <trandles> I could nominate Chris Layton from ORNL to talk about their "birthright" computing concept and how they manage it. The idea is any scientist at ORNL is entitled to some amount of free quota in their cloud to use as they see fit.
21:33:24 <oneswig> Nice name for it.  Do you get more quota if you have more kids?
21:33:30 <trandles> There's operational stuff there, like oversubscription, allowing folks to combine their birthright allocations to create larger projects, temporarily granting people an increased quota, and clawing back unused allocations.
21:35:10 <trandles> I think it's a neat concept. It gives people the ability to get work done at a scale larger than their laptop at zero cost to them. It also allows people to take risks and try things they might not be willing to try to write into a grant proposal.
21:35:21 <martial> sounds very interesting indeed
21:35:58 <martial> everyone gets their own project?
21:37:40 <trandles> yes, everyone gets their own project
21:40:11 <oneswig> This was a use case that CERN articulated, and led to some feature requests around hierarchical projects and nested quotas.
21:41:30 <martial> exactly what I was wondering: how do "users" collaborate on a common task, quota wise this must be add some complexity?
21:50:30 <oneswig> Anything more to add today?
21:51:03 <martial> not much beside
21:51:41 <martial> we have posted the link to the Etherpad on Slack, please follow up there if you intend to do a lighting talk, we do not yet have an idea of time
21:51:57 <martial> (given that we will be running on Google Meet again, so I can start/end when I want)
21:53:18 <oneswig> OK, thanks martial
21:53:31 <oneswig> Let's close it there
21:53:34 <oneswig> #endmeeting