21:00:26 <oneswig> #startmeeting scientific-sig
21:00:27 <openstack> Meeting started Tue Mar 16 21:00:26 2021 UTC and is due to finish in 60 minutes.  The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:28 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:30 <openstack> The meeting name has been set to 'scientific_sig'
21:00:47 <oneswig> Greetings, Americas region
21:01:55 <martial_> greetings Europe :)
21:02:13 <martial_> oops wait
21:02:19 <oneswig> No agenda has been set for today, as yet - was hoping to follow on from last week's discussion on JupyterHub but we don't have a guest speaker
21:02:28 <oneswig> Hi martial_, greetings
21:02:32 <oneswig> #chair martial_
21:02:33 <openstack> Current chairs: martial_ oneswig
21:02:47 <ewimmer_> Hi
21:02:48 <martial_> actually did this end up being recorded ?
21:03:11 <trandles> Hi all
21:03:34 <oneswig> Not this time martial_ - I think a session could be done and recorded on this topic in future (PTG) but it was short notice to get everything presentable
21:03:38 <oneswig> hi ewimmer_ trandles
21:04:38 <oneswig> verdurin has spoken today to connections at Compute Canada to follow up on Magic Castle and I think they'd like to join a session, but with a little more notice
21:06:39 <oneswig> martial_: what were you doing wrt P2302?  Is that the federation standard?
21:07:00 <martial_> P2302 is the IEEE Federation standard indeed
21:07:11 <martial_> we are working at multiple levels
21:07:19 <martial_> on the document and the API at this point
21:07:31 <martial_> Document is moving along nicely
21:07:48 <martial_> some differences from our CFRA work
21:08:29 <martial_> as for the API, we need to have people who have experience with OpenAPI designed RESTful interfaces have a much deeper look to understand what we are missing
21:09:10 <oneswig> What are the aims of the standard? What does it cover?
21:09:17 <janders> g'day o/
21:09:21 <martial_> this mogning was a presentation to IEEEE about the status of the future standard
21:10:07 <martial_> so "This standard defines topology, functions, and governance for cloud-to-cloud interoperability and federation. Topological elements include clouds, roots, exchanges (which mediate governance between clouds), and gateways (which mediate data exchange between clouds)."
21:10:25 <martial_> #link https://standards.ieee.org/project/2302.html
21:11:08 <martial_> in general it is about the means for clouds to have a common API to process information
21:11:23 <martial_> there is obviously a lot of pre-agreements that need to come in place
21:12:17 <martial_> hopefully this helps describe things a little more
21:12:26 <oneswig> mediated governance... what does that mean, can you give an example?
21:14:00 <martial_> two levels: pre-agreed upon levels of access + a model to exchange information about status changes
21:14:45 <martial_> we are not skipping the governance, SLAs, regulations, etc
21:15:05 <martial_> those are still around and nothing will skip those
21:15:14 <oneswig> That sounds interesting.  How to manage the federated usage of a deleted user at other clouds is a challenge.
21:16:15 <martial_> this is where each "Federation Hosting Service Operator" controls its local IdP and need to inform others of those status changes
21:17:52 <martial_> (hopefully this makes sense)
21:18:47 <martial_> today was a presentation to the IEEE of the status of the effort
21:18:51 <oneswig> If I'm running a cloud provider, is it my duty to tell all the other clouds if I delete a user?  Or is it their duty to ask me?
21:19:14 <oneswig> Could be tricky either way I guess.  Interesting problems
21:20:18 <martial_> exactly, agreements on billing are difficult
21:20:49 <martial_> I note that we do not address this, mostly the mechanisms to enable it
21:21:56 <martial_> Looks like they are saying it will be available on demand
21:22:08 <martial_> #link https://event.on24.com/eventRegistration/EventLobbyServlet?target=reg20.jsp&partnerref=mktgemail_0303&eventid=3028260&sessionid=1&key=C24196C68996F6DB7B0F4BE21A6FF2E4&regTag=1230562&V2=false&sourcepage=register
21:24:35 <martial_> what seemed to be the consensus from your jupyter talk?
21:25:44 <oneswig> People described experiences using Zero-to-Jupyterhub or Magic Castle for Jupyterhub deployment
21:26:09 <oneswig> There were some useful shared experiences around management of user authentication and data access
21:27:20 <b1airo> sorry, got caught up skimming RFP responses
21:27:32 <martial_> #chair b1airo
21:27:33 <openstack> Current chairs: b1airo martial_ oneswig
21:27:41 <martial_> welcome Blair
21:27:54 <b1airo> o/
21:28:03 <b1airo> how goes it?
21:28:07 <oneswig> b1airo: feel like waking up on Christmas day and unwrapping your presents?
21:28:17 <b1airo> a little bit :-)
21:28:55 <janders> b1airo what is it for?
21:29:44 <trandles> reading RFPs feels like Christmas Day? I don't know what kind of responses you get, but it does not feeling like that to me usually.
21:29:46 <b1airo> related... anyone used an Apollo 2000 system with Ironic? 4 nodes with a shared "Rack Consolidation Module", i.e., they all share the same physical iLO management ports
21:30:15 <b1airo> asking for a friend
21:31:03 <janders> hah
21:31:18 <janders> but each has a separate BMC IP right?
21:31:31 <oneswig> Brings up a related matter for AOB, going back to slow SSH connections from 2 weeks ago - I have an answer
21:31:35 <b1airo> janders: we're expanding our "capacity" HPC machine and building a new mixed BM/Virt multi-tenant platform
21:31:46 <b1airo> DNS?!
21:31:54 <b1airo> it's always DNS
21:32:10 <janders> b1airo sounds like some onboard BMC mini-switch
21:32:11 <oneswig> b1airo: in due course :-) and it wasn't DNS
21:32:24 <trandles> I've used Ironic with old Appro greenblades. It's kind of the same thing, one physical 1GigE connection for all 10 blades, but different MAC for each blade.
21:33:05 <b1airo> yeah, it'll be that sort of thing no doubt
21:33:05 <janders> b1airo I think I've seen this on Dell C8000 platform, don't remember having issues with this
21:33:32 <janders> I just hope they don't do something with RedFish
21:34:04 <janders> like /redfish/v1/Systems/System.Embedded.[1-10] but in some non-standard way
21:34:33 <janders> if it's standard I think it should be allright but not that I ever used them
21:34:55 <janders> and nice i/r/t capacity machine + bm/virt
21:34:58 <janders> sounds super cloudy
21:35:38 <b1airo> super cloudy indeed janders
21:35:49 <martial_> super cloudy sounds good that said :)
21:35:57 <janders> b1airo very cool! :)
21:36:11 <b1airo> a key requirement is multi-tenant bare metal as we indeed to support multiple orgs with their own managed platforms
21:36:34 <b1airo> stepping away to get coffee..
21:36:35 <janders> will you be using IB or ethernet as the main intereconnect?
21:36:58 <janders> trandles was Apollo using a reasonable Redfish implementation when you were testing it?
21:37:12 <janders> +1 for coffee, brb
21:37:25 <trandles> I've never used Apollo
21:43:58 <janders> trandles sorry misread Appro as Apollo. Clearly needing that coffee
21:44:32 <trandles> cheers!
21:47:34 <martial_> coffee power :)
21:47:48 <oneswig> I was working with some HPE blades with a recent project - the one with slow ssh connection times.
21:48:08 <oneswig> Turns out the ILO was not communicating with the chassis controller firmware
21:48:33 <oneswig> The end result was that the blades were booting clocked at 180 MHz
21:48:55 <oneswig> That's why connection times (and everything else) was inexplicably slow
21:49:21 <oneswig> doh!  Thanks HPE.
21:56:32 <b1airo> lol
21:57:03 <oneswig> I can laugh about it now...
21:57:17 <b1airo> janders: ethernet at this point
21:57:33 <b1airo> intention is to leave potential for IB though
21:59:37 <oneswig> Nearly time y'all - final comments
21:59:49 <b1airo> wish me luck!
22:00:18 <trandles> One quick note: I recruited a funded intern for the summer. He'll be working with the Ironic team on implementing kexec
22:00:29 <oneswig> Have fun with it b1airo
22:00:40 <oneswig> trandles: that's great!  Looking forward to seeing it happen
22:00:46 <martial_> cool
22:00:49 <b1airo> awesome trandles
22:01:09 <trandles> I'll drag him to one of these meetings when he gets started.
22:01:46 <oneswig> already looking forward to it
22:01:50 <oneswig> OK thanks all
22:01:57 <oneswig> #endmeeting