21:00:13 <oneswig> #startmeeting scientific-wg
21:00:14 <openstack> Meeting started Tue Mar  7 21:00:13 2017 UTC and is due to finish in 60 minutes.  The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:15 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:17 <openstack> The meeting name has been set to 'scientific_wg'
21:00:18 <armstrong> ok cool
21:00:41 <hogepodge> o/
21:00:41 <oneswig> Greetings all
21:00:52 <oneswig> #link Agenda for today https://wiki.openstack.org/wiki/Scientific_working_group#IRC_Meeting_March_7th_2017
21:01:14 <oneswig> But really it's mainly about the Forum this session
21:01:21 <flanders_> G'Day
21:01:31 <oneswig> Hi hogepodge, flanders_, armstrong and all
21:02:04 <priteau> Hello
21:02:16 <oneswig> martial will be a few minutes late (other meetings to attend to)
21:02:22 <oneswig> Hi priteau
21:02:48 <armstrong> Hello oneswig
21:02:50 <oneswig> Last week, the EMEA cohort got started on an etherpad...
21:02:52 <hogepodge> I had one item to add to the agenda if we can squeeze it in.
21:03:21 <oneswig> hogepodge: sure, up front or any other business?
21:03:49 <oneswig> #link Forum ideas brainstorming https://etherpad.openstack.org/p/BOS-UC-brainstorming-scientific-wg
21:04:18 <hogepodge> it was reprinting the openstack in science books. someone from the wg had asked for them to be reprinted for docker con, but given the expense we were hoping to defer that to sc 17 with a possible content refresh
21:05:57 <oneswig> hogepodge: I think a refresh is hoped for anyway, we started gathering data for a chapter on federation but hit a stumbling block (seems like there may be backwards compatibility issues wrt Ocata and domains - jmlowe did you find out more?)
21:06:22 <jmlowe> I did not
21:06:26 <hogepodge> sorry to bring this up out of turn on the agenda :-/
21:06:30 <oneswig> Alas, me neither
21:06:51 <oneswig> hogepodge: I wouldn't call most of our meetings strictly linear :-)
21:07:09 <martial> (sorry, another meeting that would not end)
21:07:22 <oneswig> hogepodge: what date were you thinking of as a publish deadline?
21:07:25 <oneswig> Hi martial
21:07:29 <oneswig> #chair martial
21:07:30 <openstack> Current chairs: martial oneswig
21:08:13 <oneswig> martial: people are perusing https://etherpad.openstack.org/p/BOS-UC-brainstorming-scientific-wg and we are discussing a second print run on the book
21:09:10 <hogepodge> we don't have a firm deadline, we would have to order books now for dockercon, and printing up old books means less budget for printing revised books
21:09:45 <martial> hogepodge: yes, I was the one who was wondering about the book for DockerCon
21:09:49 <oneswig> hogepodge: sounds like they are 6 months apart.
21:10:01 <martial> How many people stopped by the booth last year?
21:10:23 <armstrong> At Barcelona ?
21:10:39 <jmlowe> hey, here we go, re: federation https://bugs.launchpad.net/keystone/+bug/1642687
21:10:39 <openstack> Launchpad bug 1642687 in OpenStack Identity (keystone) "Missing domain for federated users" [Medium,Fix released] - Assigned to Ron De Rose (ronald-de-rose)
21:10:47 <martial> armstrong: sorry DockerCon 16
21:11:53 <martial> stig, we might want to also invite a couple people from the cyborg project to see us and discuss how they can help
21:12:02 <jmlowe> users no longer go in a FEDERATED domain but now have a per IdP domain
21:12:04 <martial> zhipheng is already getting in touch with blair on that
21:13:17 <oneswig> martial: you mean Cyborg team at a WG IRC meeting?
21:13:30 <martial> oneswig: I am going to see if I can come up with a small presentation on our Data Science model with benchmarking aggregation
21:13:31 <hogepodge> martial: if we want to print them up, can you circle up with me offline for number, and also justification I can take back to the marketing team?
21:13:51 <oneswig> jmlowe: what are the consequences -does this prevent federated users from being co-tenants on a common project?
21:14:07 <b1airo> Morning and apologies - wheels have come off at home this morning
21:14:17 <oneswig> b1airo: good morning
21:14:21 <oneswig> #chair b1airo
21:14:22 <openstack> Current chairs: b1airo martial oneswig
21:14:25 <martial> hogepodge: I can get in touch with you, yes. I will be at DockerCon too
21:14:52 <martial> b1airo: welcome / hopefully nothing too bad
21:15:16 <jmlowe> afaik you just need to layout your mapping a little differently
21:16:06 <oneswig> jmlowe: is this a major issue?  Something worth articulating ... in a Forum etherpad for example?
21:16:34 <martial> oneswig: no, I meant for the cyborg team to stop by for a quick BoF maybe and introduce their work to us; they are looking for input/user stories on how they can help
21:16:57 <oneswig> jmlowe: isn't federated user mapping something to keep as simple as possible, does it cause issues here?
21:16:58 <jmlowe> I think it still needs a bit more research, docs still stay you have to map to groups but I was under the impression having spoken with the keystone devs that was actually a bug fixed in Newton
21:17:47 <oneswig> jmlowe: but it's keeping you from federating - was it digital globe?
21:17:49 <martial> oneswig: zhipeng contacted me and I tried to put him in touch with b1airo, as our GPU / GPGPU expert
21:18:47 <oneswig> martial: do you have a feel for how Cyborg differs from what's being talked about with the Nova placement scheduler?  I'd be interested to see that clarified
21:19:30 <b1airo> martial: yes I got the email, haven't looked yet though
21:19:35 <martial> oneswig: not the best understanding, no, I agree; which is why I wanted to start this conversation with them
21:19:49 <martial> DFFlanders: welcome
21:20:14 <oneswig> martial: think you can clarify this with Jim G?  He's on the project isn't he?
21:20:15 <jmlowe> I'm getting my group problem fixed, read only ldap groups that were not fully populated, I need to have a fairly sophisticated mapping as all of the usage accounting depends on tenant per activity with users having roles on multiple tenants
21:20:47 <martial> oneswig: I will talk to Jim
21:20:50 <flanders_> @martial: o/
21:21:10 <jmlowe> I share a lot of my ldap backend with Chameleon
21:21:51 <oneswig> IU has fingers in many pies - is it a Chameleon site?
21:22:50 <jmlowe> It's Kate Keahey's cloud research project at UT Austin and U Chicago
21:22:54 <martial> oneswig: Jim said he will find out :)
21:23:39 <jmlowe> I believe priteau is part of that project
21:23:56 <priteau> jmlowe: Jetstream uses TACC's LDAP?
21:24:30 <martial> jmlowe: I beieve you are correct re priteau
21:24:48 <priteau> yes I am
21:25:14 <jmlowe> yes, it's insane
21:26:04 <jmlowe> I create all kinds of problems for Akhil
21:26:08 <priteau> jmlowe: we have moved away from an LDAP-backed Keystone. We now have a sync script to Keystone's MySQL DB
21:26:24 <priteau> at least in our bare-metal clouds. KVM is still using LDAP directly
21:26:38 <jmlowe> priteau: just identity, assignment is in sql
21:28:42 <priteau> We use SQL for both
21:29:07 <priteau> I can get you in touch with our admin if you want to perform the same kind of migration
21:29:15 <jmlowe> My federation goal is to do openid connect federation via globus auth and get xsede portal users mapped correctly to the tenants that their corresponding TACC users have roles on, if I can make it work I'd be happy to write it up for v2 of the book
21:29:53 <DFFlanders> ftw!
21:30:03 <martial> jmlowe: sounds great :)
21:30:07 <oneswig> jmlowe: that is EXACTLY the kind of case study that would really add value
21:33:40 <martial> what is the next item on the agenda?
21:34:09 <oneswig> some folks are working through the etherpad -
21:34:23 <oneswig> https://etherpad.openstack.org/p/BOS-UC-brainstorming-scientific-wg
21:34:50 <b1airo> Hasn't been much Forum ML chatter that I've noticed yet...
21:36:08 <oneswig> From what I've seen it's us and the Telco WG, plus Nova, Neutron who have made etherpads for brainstorming
21:38:57 <oneswig> Are there any thoughts on items that are missing from the Etherpad?
21:39:24 <DFFlanders> more to come, lots of discussions ongoing re forum as a new event.  iterations will be needed.
21:39:57 <martial> DFFlanders: yes it seems very different I must say
21:40:28 <oneswig> DFFlanders: any issues that could be laid out here that would be relevant to the Boston Cloud Declaration?
21:44:02 <trandles> I'm interested in knowing if discussions of federation & Boston Cloud Declaration are taking into consideration FedRAMP.  https://www.fedramp.gov/  It's a US gov't thing but for labs like LANL and (I presume) organizations like NIST it's an important consideration.
21:44:17 * trandles pokes martial
21:44:54 <martial> trandles: am here, but unfortunately not something I can comment on
21:46:29 <oneswig> Came up on the mailing list yesterday - http://lists.openstack.org/pipermail/openstack-operators/2017-March/012859.html
21:46:37 <oneswig> mentioned in passing
21:47:11 <trandles> gah, I missed that...thx for the link
21:47:31 <martial> trandles: I will obviously see what we can do here, but it is not something I am involved with
21:48:02 <martial> I will see who here is involved and see if they can join the converstaion
21:48:35 <trandles> martial: It's fairly new to me too.  Someone sprung it on me in a meeting yesterday.
21:49:06 <flanders_> oneswig: if we can all stay meta and observe how this forum goes (and resulting PTG), then I think we will iterate in feedback for Sydney.
21:49:44 <martial> trandles: I am aware of it, we have to use cloud systems (not on premise) that are fedramp certified, so it is on our mind, trust me :)
21:49:52 <martial> that said, it is not decided here either
21:50:23 <martial> but I will see who at NIST is the right person to get this conversation started
21:51:09 <oneswig> flanders_: ok thanks, makes sense to find a working modus operandi before driving it too hard
21:51:14 <flanders_> trandles: ping the chairs of the Boston Declaration: Khalil and Wilfred.  It is a higher level policy thing trying to get the funding agencies to agree some priority areas of which federation will be one.  I doubt they will talk technical implementation.
21:52:27 <trandles> flanders_: thx, will do
21:53:00 <martial> flanders_: agreed
21:53:52 <martial> trandles: cc me if you can, I just started reaching out here on this matter
21:54:06 <oneswig> I think the etherpad has arrived at a steady state... any more contributions?
21:54:12 <flanders_> oneswig: +1 let's use Boston to experiment and iterate the forum from there.
21:55:40 <oneswig> flanders_: point of order here, what happens next?  Is it that we fold the brainstorming sessions together and find common cause?
21:56:14 <b1airo> Still need to figure out if it is feasible to present at GTC in the middle of the Forum...
21:56:55 <b1airo> Wrong coast
21:57:02 <oneswig> They'll do a live video linkup for you, right? :-)
21:57:11 <flanders_> oneswig: the UC will make announcements on the ML
21:57:37 <oneswig> ah ok, will keep an eye out for that
21:57:45 <flanders_> In short, I'm not sure and will bring this up to UC chairs
21:57:55 <martial> flanders_: ok, thanks
21:58:03 <flanders_> Keep the questions coming!
21:58:16 <oneswig> Time's running low however
21:58:36 <armstrong> when is launchpad switching to storyboard?
21:58:55 <flanders_> I'd like the scientific-wg to be the first to know it's features are being delivered (via the PTG) for its users.
21:59:23 <b1airo> That would be nice!
21:59:41 <oneswig> A highly satisfying result indeed
21:59:52 <flanders_> Once list of forum activities are set let's think about which PTLs to invite to our sessions <-- future agenda item plz
21:59:52 <b1airo> In other news we finished our DC move!
22:00:06 <martial> b1airo: congratulations :)
22:00:12 <oneswig> armstrong: not sure - saw something about it happening but no dates nor further details...
22:00:34 <armstrong> oneswig ok thanks
22:00:38 <oneswig> b1airo: good job.  Can't wait to hear how you're going to top that one
22:00:41 <flanders_> b1airo: bonza! ;)
22:01:01 <flanders_> Are you available for squash now b1airo ?
22:01:03 <oneswig> Perhaps b1airo you can help Melbourne Uni with their move now?
22:01:20 <b1airo> Usually the way it goes oneswig
22:01:39 <flanders_> +1 tru dat
22:01:48 <oneswig> OK we are out of time.  Thanks all, it's a wrap
22:01:58 <oneswig> #endmeeting
22:01:59 <flanders_> L8er :)
22:02:03 <trandles> o/
22:02:23 <priteau> Bye everyone
22:02:26 <oneswig> oh.  The meeting hasn't ended.  Does that mean we can't leave?
22:02:27 <martial> thanks all, see you soon :)
22:02:29 <armstrong> bye
22:02:37 <martial> #endmeeting