17:00:02 <dtantsur> #startmeeting ironic
17:00:03 <openstack> Meeting started Mon Jun  5 17:00:02 2017 UTC and is due to finish in 60 minutes.  The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:06 <openstack> The meeting name has been set to 'ironic'
17:00:10 <mjturek> o/
17:00:12 <dtantsur> o/
17:00:19 <rloo> o/
17:00:36 <nicodemos> o/
17:00:37 <TheJulia> o/
17:00:38 <stendulker> o/
17:00:39 <NobodyCam> o/
17:00:57 <dtantsur> hi everyone! welcome to our small ironic part...serious meeting :)
17:01:20 <fellypefca> o/
17:01:22 <NobodyCam> :)
17:01:24 <dtantsur> our very light agenda is as usual at
17:01:26 <dtantsur> #link https://wiki.openstack.org/wiki/Meetings/Ironic
17:01:27 <mjturek> :)
17:02:09 <izumi777> o/
17:02:23 <baha> o/
17:02:38 <dtantsur> #topic Announcements / Reminder
17:03:05 <dtantsur> we've had an IPA intermediary release
17:03:10 <crushil> \o
17:03:14 <rpioso> o/
17:03:26 <dtantsur> our stable releases are still pending with 2x +2, I guess I need to ping someone (ttx?) to proceed with them
17:04:16 <dtantsur> finally, this is the weak of Pike-2. I don't plan on any releases of ironic, just something to keep in mind.
17:04:33 <rama_y> o/
17:04:37 <ricardoas> o/
17:04:43 <dtantsur> anything else from anyone?
17:04:43 <krtaylor> o/
17:05:26 <TheJulia> Nothing here
17:05:37 <dtantsur> ok, moving on
17:05:47 <dtantsur> #topic Review subteam status reports (capped at ten minutes)
17:06:01 <dtantsur> #link https://etherpad.openstack.org/p/IronicWhiteBoard starting with line 94
17:06:10 <rloo> dtantsur: the critical bug, is that the ironic-lib one?
17:06:15 <dtantsur> rloo: yes
17:06:26 <rloo> sigh, and 4 more high bugs :-(
17:06:28 <dtantsur> so yeah, I've broken the ironic-lib CI, sorry for that
17:06:28 <rloo> dtantsur: thx
17:06:44 <rloo> dtantsur: you break, you pay for it :-)
17:06:48 <dtantsur> I'm working on making it building IPA from source to actually test ironic-lib changes
17:07:05 <dtantsur> and I'm still figuring out zuul-pbr-tinyipa interactions
17:07:57 <dtantsur> rloo: we're ready to review the version column again, right?
17:08:18 <rloo> dtantsur: yes, although i have a patch updating the spec. might be worth reading/agreeing with that first.
17:08:20 <xavierr> o/
17:08:36 <rloo> dtantsur: or review the patch and if you understand it w/o the spec update, that means it is good :)
17:08:42 <dtantsur> rloo: yeah, I have it starred in gerrit, just did not get to actuall reviewing
17:08:50 <mgoddard_> o/
17:09:15 <dtantsur> #info OSC driver-properties command spec https://review.openstack.org/#/c/439907/ has 2x +2 and is ready for landing
17:09:50 <dtantsur> stendulker: hi! when updating the rescue section, please change the date, so that we know *when* the status was last updated
17:09:52 <sauloaislan> o/
17:10:45 <dtantsur> sambetts: any plans to update the IPA API version code soon(ish)?
17:10:47 <rloo> dtantsur: well, maybe that really was the status for 4/10...
17:10:54 <stendulker> dtantsur: ok.
17:10:55 <dtantsur> rloo: no, there was new stuff added
17:10:59 <dtantsur> stendulker: thanks!
17:11:23 <stendulker> dtantsur: thank you,
17:11:44 <sambetts> dtantsur: yeah I need to do that I've not had a chance since the spec merges
17:11:48 <sambetts> merged*
17:12:02 <dtantsur> ack, updating the status
17:12:34 <dtantsur> sambetts: do you know if we need to do something in ironic for https://bugs.launchpad.net/ironic/+bug/1658964 ?
17:12:35 <openstack> Launchpad bug 1658964 in Ironic "[RFE] Implement neutron routed networks support in Ironic" [Wishlist,Confirmed] - Assigned to Sam Betts (sambetts)
17:12:51 <rloo> dtantsur: wrt cross-project. there was a thread last? week about install guides, do you know if we need to do anything there (agree, disagree, whatever)?
17:13:11 <dtantsur> rloo: I don't think there was action items coming out of it
17:13:27 <dtantsur> but I guess we'll have all the guides in-tree eventually
17:13:35 <sambetts> dtantsur: main thing is the physnet awareness, then I have some experimentation to do regarding neutron + agents + placement etc
17:13:37 <dtantsur> that's more or less inevitable, the exact way is still shaping
17:13:47 <dtantsur> sambetts: but that's a separate RFE, no?
17:13:47 <rloo> dtantsur: ok. to be honest, i didn't read it, all i know it was something about intree/out of tree, and putting them all under some common dir intree
17:14:21 <rloo> dtantsur: yeah, well, we have almost all our docs intree now, so better support for that would be great
17:14:36 <sambetts> dtantsur: correct, there are changes in networking-baremetal related routed networks, but not in ironic (as far as I'm aware so far)
17:14:42 <dtantsur> yeah, we need a proper user-guide, as now our user-guide is spread between install-guide and dev docs
17:14:54 <rloo> dtantsur: i think we wanted an admin guide...
17:14:57 <dtantsur> sambetts: so, can we close this RFE and remove it from the statuses?
17:15:11 <dtantsur> rloo: right
17:15:47 <mgoddard_> sambetts: did you look into what jroll was saying about physnet scheduling for routed networks not working with ironic due to the nova compute service/host mismatch?
17:16:09 <rloo> stendulker: wrt rescue, is the status updated?
17:16:11 <dtantsur> jlvillal: around? I wonder if we still plan on splitting the tempest plugin. if so, let's get a plan on doing it (maybe as part of today's discussions)
17:16:19 <rloo> dtantsur: jlvillal still on vacation
17:16:28 <dtantsur> ah, ok
17:16:54 <rloo> dtantsur: he is back next week
17:17:02 <sambetts> mgoddard_: yeah, I had loads of disucssions about it at the summit, and it turns out the aggregates that are used for routed networks are placement service aggregates which != nova host aggregates, and the placement aggregate operate per baremetal node
17:17:06 <dtantsur> cool, thanks rloo
17:17:17 <sambetts> mgoddard_: not just per nova compute
17:17:21 <sambetts> mgoddard_: so \o/
17:17:31 * dtantsur is super confused by all these aggregates and the placement changes
17:17:37 <mgoddard_> sambetts: ok, that's good to know
17:17:43 <sambetts> dtantsur: how are we tracking networking-baremetal work? and there is some stuff I'd like to discuss about whether it lives in networking-baremetal or in ironic as a periodic task
17:17:50 <dtantsur> sambetts: I wonder if we (as in "you" ;) can have some knowledge sharing about recent nova stuff
17:18:16 <dtantsur> sambetts: we are not particularly tracking it so far. mostly because less than a half of people here understand it well
17:18:25 * TheJulia looks at the clock
17:18:29 <dtantsur> suggestions are welcome, I highly encourage someone to take initiative
17:18:33 <sambetts> dtantsur: sure, jaypipes gave some great talks at the summit to bring people up to speed too, I wonder if we can get hold of those recordings/slides to help people understand
17:18:40 <dtantsur> TheJulia: you're right; but these are good discussions
17:18:43 <dtantsur> :)
17:18:48 <TheJulia> dtantsur: absolutely ;)
17:18:54 <dtantsur> sambetts: mind asking him please?
17:19:11 <dtantsur> we're out of time indeed, let's reserve "how to track networking-baremetal work" to the discussions
17:19:13 <sambetts> dtantsur: I'll ping him and see if we can dig them out
17:19:15 <TheJulia> dtantsur: and I feel like we need a map, or a kanban board, or something for all the networking stuff at this point.
17:19:28 * TheJulia will settle for ether pad too
17:19:35 <dtantsur> yeah
17:19:40 <dtantsur> everyone done with the statuses?
17:19:52 <dtantsur> we need to agree on priorities, and then we can jump back into discussing cool stuff :)
17:20:18 * dtantsur will assume "yes" really soon
17:20:37 * rloo done except for rescue but stendulker hasn't replied
17:20:42 <dtantsur> #topic Deciding on priorities for the coming week
17:21:00 <dtantsur> so, we've landed two documentation patches, thanks all for reviews
17:21:24 <dtantsur> the OSC spec is really, really close
17:21:37 <NobodyCam> great work on the doc patches!
17:21:42 <dtantsur> ++
17:21:44 <dtantsur> should we take the version column back?
17:21:55 <rloo> ++ :)
17:22:01 <TheJulia> so, re BFV priority, I completely missed the meeting last week, and I guess I'm the designated victim (wait, is that right?) to update that patch this week, but I wouldn't have it as #1 at the moment.
17:22:14 <rloo> dtantsur: do you think the spec update should be approved first?
17:22:30 <dtantsur> mmm, we can do it in parallel, hopefully
17:22:34 <dtantsur> rloo: do you have a link handy?
17:22:50 <rloo> https://review.openstack.org/#/c/469940/
17:23:15 <dtantsur> TheJulia: to be honest, I'm worried about BFV. we have so much stuff ahead. I'd keep pushing on it, if we still hope for Pike.
17:23:18 <dtantsur> thanks rloo
17:23:36 <rloo> when's the nova deadline for feature code?
17:23:47 <TheJulia> dtantsur: likewise :(
17:24:00 <dtantsur> rloo: Pike-3 IIRC
17:24:09 <mjturek> TheJulia: would it help if I took over the wire-in patch or you good?
17:24:39 <TheJulia> mjturek: if you have the bandwidth, that would help me a lot.
17:24:40 <rloo> dtantsur: week of july 24. we have a bit over 1 month.
17:24:48 <mjturek> TheJulia: will do
17:25:04 <dtantsur> rloo: yeah, which leaves us really no time, unless we land the whole ironic part in 2-3 weeks..
17:25:13 <rloo> dtantsur: well, we have maybe at most 1 month. ironic code for bfv needs to work before nova will look at the nova patch and that needs to land by week of july 24.
17:25:18 <dtantsur> I mean, my optimistic hope is to finish ironic and ironicclient bits
17:25:22 <stendulker> rloo: lost connection. rescue status is updated.
17:25:28 <rloo> thx stendulker
17:25:40 <rloo> TheJulia, mjturek: you think it is doable? ^^
17:25:48 <dtantsur> mjturek: thanks, that's awesome! we're getting good pace, let's keep it :)
17:26:13 <mjturek> :)
17:26:26 <TheJulia> rloo: if we have sufficient review bandwidth and get revisions updated quickly, I think so.
17:26:43 <dtantsur> derekh (on holidays today) seems to get some progress with testing
17:26:50 <rloo> TheJulia: thx.
17:27:06 <mjturek> rloo: I think so. I know that derekh was able to boot from volume with the current code
17:27:09 <dtantsur> "have sufficient review bandwidth" is the reason I want it on the priority list
17:27:22 <TheJulia> mjturek: that is music to my ears
17:27:28 <mjturek> :)
17:28:14 <rloo> so is bfv the 'highest' priority cuz of nova? are there other features that need nova changes too?
17:28:16 <TheJulia> dtantsur: makes sense, I really can't be one to review my own code, but I can tackle other stuff :)
17:28:33 <dtantsur> rloo: it matches the current priorities order
17:28:42 <rloo> the network stuff worries me too. maybe we should put one of those patches for this week.
17:28:44 <dtantsur> rloo: but I don't care about relative order too much, to be honest
17:28:58 <dtantsur> rloo: I'm all for it, given that the OSC spec is nearly done.
17:29:03 <dtantsur> does anyone has a patch in mind for ^^^?
17:29:21 * SotK will keep it similar
17:29:31 <SotK> sorry, wrong window :/
17:29:34 <rloo> dtantsur: i think the pyhsical network awareness stuff is needed
17:29:52 <rloo> dtantsur: https://review.openstack.org/#/c/461301/ is first in series
17:30:08 <rloo> sambetts: ^^ is that your understanding?
17:30:11 <rloo> mgoddard_: ^^ ?
17:30:26 <sambetts> +1 that would be a great series to start, I think its a pretty straight forward model and API change for ports
17:30:51 <mgoddard_> rloo: that's correct
17:30:59 <dtantsur> ok, it's there. how does the list look now?
17:31:16 <dtantsur> are we ok with landing the API bits even if they're not wired in?
17:31:37 <sambetts> not sure what you mean not wired in?
17:32:05 <dtantsur> I'm not sure either :) what will happen if users put stuff to the new fields/tables?
17:32:08 * TheJulia wonders which patches exactly
17:32:30 <dtantsur> TheJulia: I'm talking about https://review.openstack.org/#/c/461301/
17:32:38 <sambetts> dtantsur: I believe it'll influence how their neutron ports are mapped onto phyiscal interfaces
17:33:00 <rloo> dtantsur: that patch isn't API; it is db change I think
17:33:09 <dtantsur> ooops
17:33:12 * dtantsur is stupid
17:33:17 <dtantsur> thanks rloo, you're right
17:33:21 <sambetts> yup, its a classic 3 part patch, DB first, Conductor, then API
17:33:27 <mgoddard_> until later patches are merged there should be no effects, even if the user modifies the DB
17:33:30 <dtantsur> ok, any more objections to the list?
17:33:34 <rloo> dtantsur: i think sambetts mentioned API, but I think he was referring to the entire feature :-)
17:33:43 <dtantsur> ack :)
17:33:48 <TheJulia> dtantsur: none
17:34:08 <rloo> list looks good to me
17:34:12 <dtantsur> ok, let's move on?
17:34:16 <mgoddard_> this one modifies the VIF attachment algorithm, which is where things change: https://review.openstack.org/#/c/470406/1
17:34:31 <dtantsur> thanks mgoddard_
17:34:33 <dtantsur> #topic Open discussion
17:34:57 <dtantsur> sambetts: do you want to discuss what we do with networking-baremetal?
17:35:11 <dtantsur> I personally like what the BFV subteam does with their etherpad and a separate meeting
17:35:40 <sambetts> yeah I think we can start that stuff again to cover the new features
17:35:50 <sambetts> not sure if we need a meeting or not yet
17:36:05 <TheJulia> Well, we kind of already have a meeting slot, it is just not actively meeting :)
17:36:10 <sambetts> my main question is how are we managing the code that is living in networking-baremetal
17:36:16 <dtantsur> TheJulia: for networking?
17:36:25 <TheJulia> dtantsur: yeah, neutron integration specifically
17:36:34 <dtantsur> mmm, we should use it then
17:36:43 <dtantsur> sambetts: wdym by "managing" here?
17:36:55 <NobodyCam> dtantsur: saw the comment on my RFE for not adding ports, Thank you. and just a quick thought on none vs noadd I picked NoAdd so it would be confused with the python None in a conf file.
17:36:58 <rloo> TheJulia: we had that meeting; do we still have the slot? It used to be before this meeting i think.
17:37:01 <sambetts> RFE's reviews, releases etc etc
17:37:21 <dtantsur> NobodyCam: good call, hmmm. "disabled"?
17:37:25 <TheJulia> rloo: we still have it, it was still on eavesdrop last time I looked a few weeks ago
17:37:36 <NobodyCam> dtantsur: ++ that works!
17:37:37 <rloo> TheJulia: good
17:37:47 <TheJulia> http://eavesdrop.openstack.org/#Ironic/neutron_Integration_team_meeting
17:37:47 <dtantsur> sambetts: releases are simple: the release team is doing them with the help of the release liaison and the PTL
17:37:59 <rloo> TheJulia: we should probably put it/link back on our meetings page.
17:38:14 <dtantsur> TheJulia: I have internal conflicts for this slot :(
17:38:23 <dtantsur> well, people can meet without me obvoiusly
17:38:36 <TheJulia> I am happy to attend, as long as I'm not on a plane
17:38:46 <rloo> sukhdev was chairing that meeting. do we want neutron folks there, or is this to get our/ironic act together?
17:38:46 <dtantsur> sambetts: RFE reviews, I think, should happen in this meeting. we even have a topic for that.
17:39:06 <TheJulia> rloo: I think it is get our act together
17:39:07 <sambetts> dtantsur: do they go into launchpad/ironic?
17:39:12 <sambetts> or their own project?
17:39:16 <TheJulia> rloo: we can always adjust the meeting title
17:39:20 <dtantsur> sambetts: we tend to have launchpad projects per project
17:39:24 <rloo> TheJulia: in that case, we should pick a time where dtantsur can attend.
17:39:40 <dtantsur> sambetts: it depends on the project-config configuration for this repo. we can check after the meeting
17:40:01 <rloo> sambetts: where does networking-baremetal live? (under neutron, under ironic, under ?)
17:40:12 <sambetts> rloo: under ironic
17:40:30 <sambetts> i think (although thats another question I had)
17:40:34 <rloo> sambetts: ok, so we treat it like we treat all/most of the other projects under ironic
17:40:55 <rloo> sambetts: ha ha, so we're not sure. that's probably the first question to answer then.
17:41:26 <sambetts> I think it should live under Ironic because I believe there will be things in there soon that might become hard requirements to make ironic+neutron work
17:41:29 <rloo> sambetts: are we the only project to ahve this? I thought nova or some other project had a similar thing. networking-baremetal is the plugin for neutron, right?
17:41:57 <dtantsur> sambetts, rloo, it's under ironic: https://github.com/openstack/governance/blob/771e492a997c528f60019d148f9cf824dc294eae/reference/projects.yaml#L1749
17:41:57 <sambetts> rloo: yeah, but neutrons model for plugins is they don't live in neutron
17:42:05 <sambetts> dtantsur: oh awesome :)
17:42:12 <sambetts> rloo: s/in/under
17:42:16 <rloo> sambetts: so you answered my question. it lives in ironic then :-)
17:42:19 <dtantsur> so I'm its PTL for now, and you can throw questions at me :)
17:43:28 <dtantsur> https://github.com/openstack-infra/project-config/blob/master/gerrit/projects.yaml#L5136-L5138 is its configuration
17:43:45 <sambetts> looks like we have a separate launchpad for it already
17:43:48 <dtantsur> as I don't see "groups" configuration, then a separate launchpad project should be created
17:43:49 <sambetts> https://launchpad.net/networking-baremetal
17:43:50 <dtantsur> aha
17:44:05 <sambetts> it appears that we need to change its ownership
17:44:14 <dtantsur> sambetts: we need to ask vsaienk0 to move it under ironic-drivers (both "driver" and "maintainer")
17:44:56 <dtantsur> sambetts, rloo: does it clarify something?
17:45:02 <rloo> yup
17:45:33 <sambetts> yup sounds good, so we treat it similar to ironic-inspector and RFE's go into the separate project
17:45:36 <rloo> i like that it provides deep ironic/neutron integration, not shallow, but deep :-)
17:45:46 <sambetts> hahaha
17:46:02 <sambetts> deep dark intergration
17:46:13 <sambetts> black magic if you will
17:46:16 <dtantsur> sambetts: yes. and we have a core subteam, like in ironic-inspector
17:46:34 <dtantsur> this will help with reviews
17:46:41 <dtantsur> s/we have/we can have/
17:46:56 <TheJulia> sambetts: ironic magic ;)
17:47:04 <sambetts> I guess it'll have its own section on the whiteboard to o
17:47:09 <dtantsur> yes
17:47:11 <rloo> dtantsur: the rfes for networking-baremetal may affect ironic code. how do we coordinate that?
17:47:16 <dtantsur> actually, we have the same situation with sushy and virtualbmc
17:47:39 <sambetts> my guess it the same as with insepction ones we create them in inspector with multiple projects affect right?
17:47:42 <dtantsur> rloo: what we do with inspector is to have a spec againt ironic, if it covers both projects, against inspector otherwise
17:47:57 <rloo> dtantsur: ok, that makes sense.
17:47:58 <dtantsur> this is about specs. for RFEs you can add many projects to them
17:48:08 <sambetts> got it
17:48:09 <dtantsur> and then approve the overall request
17:48:41 <sambetts> we don't need a networking-baremetal-specs repo at the momement then?
17:48:56 <dtantsur> sambetts: I don't see a pressing need for that
17:49:12 <dtantsur> we started having it for ironic-inspector when we got complex features that did not touch ironic essentially
17:49:15 <dtantsur> like inspector HA
17:49:33 <sambetts> yeah, I think we'll see how the project grows first
17:50:30 <dtantsur> with this in mind: please add networking-baremetal, virtualbmc and sushy to your review dashboards, depending on your area of expertise
17:50:56 <dtantsur> (and of course ironic-inspector and bifrost, but things look happier there for now)
17:51:16 <dtantsur> anything else to discuss?
17:51:54 <dtantsur> sambetts: wanna take an action item to establish and etherpad/trello/whatever and revive the meeting?
17:52:35 <sambetts> dtantsur: sure
17:52:38 <dtantsur> thanks!
17:52:54 <dtantsur> if nobody has anything to bring up, I'm ready to finish a bit earlier :)
17:53:25 <TheJulia> Thanks!
17:53:35 <dtantsur> thanks everyone, a good meeting
17:53:42 <dtantsur> #endmeeting ironic