15:00:49 <TheJulia> #startmeeting ironic
15:00:49 <opendevmeet> Meeting started Mon Jun 14 15:00:49 2021 UTC and is due to finish in 60 minutes.  The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:49 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:49 <opendevmeet> The meeting name has been set to 'ironic'
15:00:50 <TheJulia> o/
15:00:52 <dtantsur> o/
15:00:57 <ajya> o/
15:00:57 <rpittau> o/
15:01:03 <rpioso> \o
15:01:08 <stendulker> o/
15:01:17 <TheJulia> Greetings everyone, and welcome to our weekly meeting!
15:01:27 <arne_wiebalck> o/
15:01:32 <TheJulia> Our agenda can be found on the wiki
15:01:34 <TheJulia> #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting
15:02:16 <TheJulia> #topic Announcements / Reminders
15:02:38 <TheJulia> I don't believe we have anything to announce or remind ourselves of this week. Does anyone have anything to raise this week?
15:03:20 * TheJulia hears the crickets appear and gives it a few more moments
15:03:53 <TheJulia> Looks like we have no action items from last week
15:04:13 <TheJulia> I guess we can proceed to subteam status reports then
15:04:55 <dtantsur> yep
15:05:00 <TheJulia> #topic Review subteam status reports
15:05:05 <TheJulia> #link https://etherpad.openstack.org/p/IronicWhiteBoard
15:06:52 <TheJulia> arne_wiebalck: By chance have you discussed node error history with kaifeng or are we basically at no update this week?
15:07:03 <arne_wiebalck> sorry, no update
15:07:22 <arne_wiebalck> I will try to pick this up
15:07:26 <TheJulia> ack
15:07:37 <TheJulia> iurygregory_, rpittau: any update on privsep?
15:07:52 <JayF> I'll note it's not technically a subteam, but zer0c00l put up a PR for Anaconda driver docs. I added a couple of folks as reviewers and marked it as a review prio.
15:07:52 * TheJulia senses iurygregory_ is out today
15:08:05 <rpittau> no updates
15:08:11 <TheJulia> JayF: line 75 :)
15:08:23 <rpittau> not sure we want to have the discussion today since iury's not here
15:08:27 * JayF is multitasking in a downstream meeting and may not have checked the whiteboard
15:08:30 <TheJulia> rpittau: ack
15:08:34 <TheJulia> JayF: ack
15:08:38 <dtantsur> yeah, I think Iury took a day off
15:08:57 <dtantsur> TL;DR I was concerned about launching a subprocess in IPA
15:09:06 <dtantsur> given the "fun" around eventlet
15:09:19 <dtantsur> but yeah, we can wait for him to come back (although I won't be here next Monday.. so ironic)
15:09:40 <TheJulia> dtantsur: ack, okay
15:09:47 <TheJulia> hmm
15:10:58 <TheJulia> Well, Onward to priorities for the coming week?
15:11:19 <dtantsur> yep
15:11:35 <TheJulia> #topic Deciding on priorites for the coming week
15:11:37 <TheJulia> #link https://review.opendev.org/q/hashtag:ironic-week-prio+status:open
15:12:04 <dtantsur> I have two driver_info clean-ups to add: https://review.opendev.org/c/openstack/ironic/+/796287 https://review.opendev.org/c/openstack/ironic/+/796289 cc stendulker
15:12:06 <TheJulia> I went through about an hour ago and I didn't see much that could be added to the list at present. Does anyone have anything they would like to see on the list?
15:12:23 <stendulker> dtantsur: Will have a look
15:12:33 <dtantsur> thanks! I'm asking you because they touch iLO
15:12:48 <stendulker> dtantsur: np. Thank you
15:13:01 <dtantsur> any objections to adding these 2 patches?
15:13:13 <TheJulia> none, they have been added
15:13:19 <TheJulia> Does anyone have anything else this week?
15:13:20 <dtantsur> cool thx
15:13:22 <arne_wiebalck> TheJulia: I guess your -1 on https://review.opendev.org/c/openstack/ironic/+/768009 can be removed, no? This may prevent further reviews.
15:13:56 <TheJulia> arne_wiebalck: likely needs to be rebased, fwiw
15:13:56 <dtantsur> can I add baremetal-operator PRs? :D (jk)
15:14:07 <rpittau> lol
15:14:12 <TheJulia> oh my
15:14:25 <TheJulia> so onward to the baremetal sig?
15:14:32 <dtantsur> ++
15:14:47 <rpittau> let's
15:15:08 <TheJulia> #topic Baremetal SIG
15:15:17 <arne_wiebalck> dtantsur's presentation on the IPAB is uploaded to the usual youtube channel
15:15:24 <TheJulia> Awesome!
15:15:30 <dtantsur> #link https://www.youtube.com/watch?v=1L1Ld7skgDw IPA-builder introduction
15:15:33 <arne_wiebalck> thanks again, dtantsur !
15:15:35 <TheJulia> has a blog entry been posted to ironicbaremetal.org ?
15:15:37 <rpittau> nice
15:15:41 <dtantsur> I haven't posted anything
15:16:09 <arne_wiebalck> TheJulia: should we make this the routine for new presentations?
15:16:16 <TheJulia> Okay, If I get a few minutes I'll try to do that
15:16:22 <arne_wiebalck> ty
15:16:33 <arne_wiebalck> I don't think there is anything else for the SIG.
15:16:57 <TheJulia> arne_wiebalck: I think it would be good, but we need to make it a community habit to update the blog with $new stuff that is not just the videos from the SIG
15:17:14 <arne_wiebalck> TheJulia: ++
15:17:15 <TheJulia> Well, in that case, time for Open Discussion
15:17:18 <TheJulia> #topic Open Discussion
15:17:24 <TheJulia> arne_wiebalck: looks like you had a topic :)
15:17:48 <arne_wiebalck> yes, it is about network burn-in
15:17:59 <arne_wiebalck> cpu, mem, disk are merged
15:18:13 <arne_wiebalck> network is slightly more complicated as we need two nodes
15:18:24 <arne_wiebalck> at CERN, we use dynamic pairing for this
15:18:30 <TheJulia> Interesting
15:18:42 <arne_wiebalck> initially with a message queue, atm with a shared file
15:18:55 <arne_wiebalck> for upstream I was thinking to use tooz/zookeeper
15:19:00 <arne_wiebalck> for the match making
15:19:12 <TheJulia> I know, for a while I had a semi-dedicated iperf target for such things, But pairing can make sense from a "can two nodes talk really fast to each other"
15:19:20 <arne_wiebalck> the first iteration may use static pairs in driver_info
15:19:40 <arne_wiebalck> just wanted to check if the overall plan sounds ok
15:19:55 <dtantsur> why not use another fixed server?
15:20:10 <arne_wiebalck> you mean like the controller?
15:20:16 <arne_wiebalck> bandwidth mostly
15:20:29 <arne_wiebalck> imagine 100 servers trying to connect
15:20:33 <TheJulia> you also want to check cross-node bandwidth between the racks potentially
15:20:36 <TheJulia> oh, and yeah
15:20:44 <TheJulia> if your doing 100 at a time your going to have saturation issues
15:20:53 <arne_wiebalck> dtantsur: it is an option which the static approach would cover as well
15:21:18 <TheJulia> I think the idea makes sense and the approach of simple static and then evolve it makes sense as long as we don't explicitly add requirements for new services as a hard operating requirement
15:21:26 <arne_wiebalck> dtantsur: but it is not only connectivity, but also if we can go full steam
15:21:57 <arne_wiebalck> TheJulia: this is why I thought tooz as we (can) use this already
15:22:19 <JayF> It's going to  be tough to get creds for something like that down into the agent, isn't it?
15:22:23 <arne_wiebalck> TheJulia: for leader election in inspector, for instance
15:22:37 <TheJulia> Last time I used iperf, it had no concept of credentials
15:22:47 <JayF> I'm saying if tooz is going to be used
15:22:53 <TheJulia> ahh
15:23:02 <JayF> there's not any cleaning steps, afaik, which coordinate between ironic and ipa for the step
15:23:05 <TheJulia> well, it could be determied conductor side and fed to the agent
15:23:15 <JayF> so if you're locking for access to the network server, you're going to need creds to do that locking in IPA
15:23:26 <JayF> or come up with a new method for IPA clean steps to coordinate with conductor
15:23:38 <JayF> which is probably a lot more complexity than desired for this problem?
15:23:43 <TheJulia> I don't think tooz is necessarilly for locking
15:23:53 <TheJulia> at least, that is not my perception, at least directly
15:23:59 <JayF> Heh. Clean steps [lock network server] [ network burn in ] [unlock network server]
15:24:00 <TheJulia> but getting creds in, is indeed a problem
15:24:00 <arne_wiebalck> TheJulia: it can be, I think
15:24:04 <TheJulia> and then communication flows
15:24:06 <JayF> where 2 of them are in the conductor and one is in IPA
15:24:13 <JayF> and you'd only need to ensure the priorities lined up
15:24:19 * JayF not sure he loves that but it would likely work
15:24:30 <dtantsur> it seems like you're enabling someone with IPA access to lock IPA on another server?
15:24:42 <dtantsur> (not sure if it's a concern or not)
15:25:02 <TheJulia> Lets take this path, simple start, see where that leaves us, and then iterate from there?
15:25:06 <JayF> I'm concerned about this feature in general which is why I'm trying to hash it out :) see if it's less complex than I thought itw as at first glance
15:25:10 <arne_wiebalck> sounds good to
15:25:12 <arne_wiebalck> me
15:25:12 <TheJulia> so static operator supplied info, and then see if we can do automagic
15:25:17 <arne_wiebalck> static works as well
15:25:31 <arne_wiebalck> it is just not so nice if you have broken servers
15:25:42 <arne_wiebalck> ok, thanks everyone!
15:26:01 <TheJulia> Does anyone else have anything for Open Discussion?
15:27:21 <dtantsur> not me
15:27:49 <lmcgann> Id just like to remind everybody about our work on the security interface spec: https://review.opendev.org/c/openstack/ironic-specs/+/576718/27/specs/approved/security-interface.rst
15:27:55 <TheJulia> In that case, the last step of our meeting is next
15:28:27 <TheJulia> lmcgann: perhaps a few words to express why it is important to those in attendance of today's meeting?
15:31:09 <TheJulia> I know I'm kind of putting you on the spot, but it helps people relate and understand why
15:32:41 <TheJulia> Anyway, I guess we can proceed with our agenda
15:32:47 <TheJulia> #topic Who is going to run the next meeting?
15:32:56 <TheJulia> Any volunteers ?
15:33:02 <dtantsur> I cannot - will be out for a language exam (doh)
15:33:10 <TheJulia> eek
15:33:17 <arne_wiebalck> German ... should be fun!
15:33:18 <TheJulia> should we switch IRC to German for the week?
15:33:34 <TheJulia> Granted, it will mostly be machine translated for many of us
15:33:39 <arne_wiebalck> only for dtantsur maybe
15:33:44 <TheJulia> hehe
15:34:14 <TheJulia> Well, I can do it next week, I guess
15:34:16 <rpittau> I think I can run the next meeting
15:34:28 <TheJulia> rpittau: you sure?
15:34:31 <lmcgann> sorry for the delay TheJulia, Im trying to think of a good way to describe the spec
15:34:33 <rpittau> yeah :)
15:34:34 <JayF> dtantsur: Wie werde ich all diese Server bereitstellen?
15:34:46 <dtantsur> :D
15:34:49 <rpittau> but it will be in Italian :P
15:35:03 <arne_wiebalck> molto bene!
15:35:07 <rpittau> lol
15:35:12 <TheJulia> lmcgann: simplest phrases or expressions generally help, or even the desired end result
15:35:37 <TheJulia> rpittau: in Italian, ugh.
15:35:50 <NobodyCam> Good Morning Ironic Folks, Happy Monday!
15:35:56 * TheJulia should actually like... learn some about Italian explicitly given Latin experience
15:36:35 <dtantsur> I will use morse alphabet. using the word "pizza" or absence of it.
15:36:37 <TheJulia> rpittau: Okay, I guess your running the meeting next week if nothing comes up
15:36:48 <rpittau> cool :)
15:37:07 <rpittau> dtantsur: that's a way :D
15:37:23 <TheJulia> dtantsur: so do we convey everything with pizzapizza or pizza and translate that to dots and dashed and then that to words?!
15:37:35 <TheJulia> dtantsur: or Norse alphabet?
15:38:01 <dtantsur> for Norse alphabet we may need hjensas to run the meeting
15:38:11 <TheJulia> dtantsur: this is true...
15:38:26 <TheJulia> Anyway, Thanks everyone, have a wonderful week!
15:38:54 <arne_wiebalck> Thanks TheJulia !
15:38:59 <rpittau> thanks!
15:39:04 <TheJulia> #endmeeting