17:00:03 <dtantsur> #startmeeting ironic
17:00:04 <openstack> Meeting started Mon Jul 17 17:00:03 2017 UTC and is due to finish in 60 minutes.  The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:07 <openstack> The meeting name has been set to 'ironic'
17:00:14 <mgoddard_> o/
17:00:15 <nicodemos> o/
17:00:15 <TheJulia> o/
17:00:17 <mjturek> o/
17:00:22 <ricardoas1> o/
17:00:24 <kaifeng> o/
17:00:26 <dtantsur> hello everyone!
17:00:31 <etingof> o/
17:00:35 <rloo_> o/
17:00:39 <rama_y> o/
17:00:53 <crushil> \o
17:00:55 <vdrok> o/
17:01:07 <fellypefca> o/
17:01:08 <stendulker> o/
17:01:09 <rpioso> o/
17:01:38 <dtantsur> our agenda, as usual, can be found at
17:01:40 <dtantsur> #link https://wiki.openstack.org/wiki/Meetings/Ironic
17:01:48 <dtantsur> #topic Announcements / Reminder
17:01:58 <dtantsur> #info ironic to get tag: supports-api-interoperability (https://review.openstack.org/#/c/482759/)
17:02:00 <milan> o/
17:02:08 <dtantsur> this is because we do API microversioning, mostly :)
17:02:23 <rloo_> dtantsur: yay :)
17:02:26 <dtantsur> not sure what effect it will have on us further.. nothing expected right now
17:02:54 <dtantsur> #info starting to collect early ideas for the PTG: https://etherpad.openstack.org/p/ironic-queens-ptg
17:03:16 <dtantsur> the PTG is still 2 months away, but people already mention certain potential topics to discuss.. so put them there
17:03:24 <krtaylor> 0/
17:03:30 <TheJulia> dtantsur: Two other things:  This week is non-client library release freeze, meaning sushy/virtualbmc afaik), and next week is client library release freeze.
17:03:41 <dtantsur> not virtualbmc IIRC, but sushy for sure
17:03:56 <dtantsur> we need to land what can be landed in sushy and ironic-lib by roughly Wed
17:04:01 <rloo_> dtantsur: wrt ptg -- do we need to get leads now? might be good just to get people's ideas first?
17:04:09 <TheJulia> oh, yeah, ironic-lib too :)
17:04:34 <dtantsur> rloo_: well, I want people to understand that somebody has to lead the topic, including *before* the PTG
17:05:03 <rloo_> dtantsur: i agree, but i think that if people have good ideas we should gather those, then look for leads later if we need to
17:05:05 <dtantsur> just putting your name when you submit it is fine enough for now
17:05:26 <rloo_> dtantsur: ok, that works for me. i'll update etherpad to reflect that unless someone beats me to it. thx!
17:06:08 <rloo_> wrt non-client library releases -- do we want to get the doc migration changes done for them all before the releases?
17:06:29 <TheJulia> We will, I've got the sushy repo up on my other screen and starting to work on that now
17:06:34 <dtantsur> rloo_: preferably
17:06:37 <TheJulia> at least, the basic changes
17:07:06 <dtantsur> rloo_: slightly changed the wording
17:07:15 <rloo_> dtantsur: thx!
17:07:22 <dtantsur> #info Non-client libraries final releases this week, affects sushy and ironic-lib
17:07:50 <dtantsur> anything else?
17:07:53 <rloo_> and molteniron, virtualbmc
17:08:19 <rloo_> next week is final release of client libraries, was that mentioned?
17:08:37 <rloo_> and oft string freeze which doesn't affect us mcuh
17:08:47 <dtantsur> I think virtualbmc is release type "other", but it needs double-checking
17:08:51 <rloo_> and feature freeze -- nova, which affects us
17:09:05 <dtantsur> #info Final client releases and the feature freeze (for some projects, not us) next week
17:09:13 <vdrok> one small announcement, the first patch to networking-baremetal was approved, a couple more smaller ones needed to be able to use it in gate
17:09:15 <dtantsur> now, I wonder if we need a soft feature freeze at some point
17:09:31 <dtantsur> thanks vdrok, good news!
17:09:36 <mjturek> we haven't done freezes on molteniron, it's rolling at this point. Open to changing that though
17:09:44 <rloo_> vdrok: great!
17:09:51 <dtantsur> mjturek: up to you and people using it, I guess
17:10:00 <TheJulia> mjturek: also molteniron has never been released officially before :)
17:10:06 <mjturek> fair enough!
17:10:17 <mjturek> yeaaah true lol
17:10:40 <dtantsur> ok, anything else? we can discuss a potential soft freeze in open discussions
17:11:08 <dtantsur> #topic Review subteam status reports (capped at ten minutes, or, well, as usual..)
17:11:23 <dtantsur> #link https://etherpad.openstack.org/p/IronicWhiteBoard starting with line 121
17:11:59 <rloo_> dtantsur: what's the critical ironic bug?
17:12:17 <dtantsur> I suspect the neutron one, lemme check
17:12:20 <rloo_> inspector also has a critical bug
17:12:28 <dtantsur> inspector has CI down
17:12:49 <dtantsur> oh
17:12:51 <dtantsur> #link https://bugs.launchpad.net/ironic/+bug/1699542
17:12:53 <openstack> Launchpad bug 1699542 in Ironic "ironic fails when glance-api is deployed under uwsgi as vhost" [Critical,In progress] - Assigned to Pavlo Shchelokovskyy (pshchelo)
17:12:53 <milan> yup :-/
17:13:22 <dtantsur> #link https://bugs.launchpad.net/ironic-inspector/+bug/1703804
17:13:22 <openstack> Launchpad bug 1703804 in Ironic Inspector "Grenade fails to find inspector logs" [Critical,New]
17:13:55 <rloo_> dtantsur: that ironic bug isn't critical any more? didn't you put in a fix?
17:14:20 <dtantsur> well, I guess it should be closed, lemme see
17:14:46 <vdrok> yes, there was a fix. now a refactor of keystoneauth things needed to proceed
17:15:10 <rloo_> vdrok: right. is that being tracked by another bug or this one?
17:15:14 <dtantsur> another
17:15:25 <dtantsur> #link https://bugs.launchpad.net/ironic/+bug/1699547
17:15:26 <openstack> Launchpad bug 1699547 in Ironic "[RFE] unify endpoint discovery for services by using keystoneauth adapters" [Wishlist,In progress] - Assigned to Pavlo Shchelokovskyy (pshchelo)
17:15:50 <rloo_> dtantsur: ok, this can be closed then. and phew, not critical :-)
17:15:57 * dtantsur closed
17:17:00 <rloo_> TheJulia: wrt bfv, L150 means ?
17:17:38 <TheJulia> rloo_: Sorry, incomplete thought, I was meaning that they would be in the next release of python-ironicclient,
17:17:45 <dtantsur> rloo_: we only released the python API, we need a release for CLI bits
17:17:45 <TheJulia> whatever/whenever that next release is cut
17:17:46 <milan> jlvillal has been driving the inspector grenade issue, thx jlvillal! We need infra to review the patch https://review.openstack.org/#/c/483059/
17:18:08 <rloo_> TheJulia: you mean the rest of the client patches will be in next release?
17:18:31 <TheJulia> yes
17:18:47 <rloo_> TheJulia: cuz you know, stuff in release x is also in release x+1 unless we explicitly delete them :-)
17:19:33 <rloo_> TheJulia: we should plan on another client release this week since the rest  of bfv client code should merge today.
17:20:07 <TheJulia> rloo_: heh, agreed although if dtantsur has anything he wants to see in the client before next week, we should consider that as well.
17:20:22 <dtantsur> TheJulia: physnet awareness bits?
17:20:30 <rloo_> TheJulia: wrt the nova patch for bfv -- do you think it could use more ironic eyes?
17:21:02 <rloo_> TheJulia: right, i thought it was a blocker for bfv devstack changes though, so i'm good if we do two more releases of client
17:21:18 <TheJulia> dtantsur: if we can get the api for it reviewed, sure :)
17:21:19 <dtantsur> as many as we need :)
17:21:40 <rloo_> dtantsur: yes, we should get physnet aware bits in before end of next week. i am going to try to get that in this week :)
17:21:50 * dtantsur too
17:22:04 <dtantsur> so, BFV looks very good, right?
17:22:09 <TheJulia> okay, works for me re the client releases
17:22:14 <dtantsur> I mean, missing tempest is not awesome, but still!
17:22:35 <TheJulia> it does, a couple misc patches, devstack, tempest tests, and the nova revision.
17:22:51 <dtantsur> great job, TheJulia, mjturek and everyone involved
17:23:06 <mjturek> :)
17:23:20 * rloo_ holds off the congrats until after nova gives their blessing... crosses fingers...
17:23:35 <dtantsur> true, but the ironic part is a huge achievement on its own
17:23:49 <rloo_> dtantsur: so very true but i don't want to anger the gods...
17:23:55 <dtantsur> heh, fine
17:23:57 <TheJulia> :)
17:24:15 <dtantsur> anyway, done with the statuses? we're past 10 mins and we have a nice flame war ahead ;)
17:24:34 * TheJulia wonders if we can have the flame war with drinks...
17:24:46 <rloo_> huh, what flame wars? (but yeah, ready to move on)
17:24:58 <dtantsur> rloo_: pruning the priorities :)
17:24:59 <mgoddard_> dtantsur: would be great to get physnet awareness into the client. There is still the ironic API change and tempest tests to merge first though
17:25:04 <dtantsur> right
17:25:11 <rloo_> dtantsur: no flame wars there ;)
17:25:20 <dtantsur> #topic Deciding on priorities for the coming week
17:25:51 <rloo_> lets add physnet patch
17:25:56 <dtantsur> sure
17:25:57 <TheJulia> +1
17:26:05 <sambetts> +1
17:26:52 <dtantsur> ok, the docs work stays, because it has to
17:27:02 <dtantsur> BFV, I guess, stays, because we're SOOO close
17:27:13 <dtantsur> rolling upgrades - we have to get rolling (see?) with them
17:27:22 <dtantsur> mmm, any other wise thoughts?
17:27:33 <rloo_> i think those are good weekly priorities
17:27:39 <TheJulia> agreed
17:27:50 <dtantsur> yeah, the list looks big enough to me, given also the inspector CI problem
17:28:21 <dtantsur> moving on? any objections?
17:28:25 <TheJulia> no objections
17:28:40 <rloo_> + move on
17:28:47 <dtantsur> #topic Prune Pike priorities
17:28:56 <dtantsur> a sad, but necessary, thing to do
17:29:18 <rloo_> dtantsur: would be useful to decide on the soft freeze
17:29:23 <dtantsur> rloo_: I guess the plan is to clean the priorities list of things that are clearly not making it, right?
17:29:50 <rloo_> dtantsur: right. clean or at least downgrade. so it represents what we reasonably be able to get done.
17:29:51 <dtantsur> we can as well. by soft freeze I meant only merging 1. priorities, 2. small vendor-contained features, 3. bug fixes and docs
17:30:02 <rloo_> dtantsur: i don't think there will be any surprises but...
17:30:24 <rloo_> oh, soft freeze includes bug fixes? ok
17:30:41 <rloo_> rc1 targetweek is Aug 7
17:30:41 <dtantsur> is it a surprise? ;)
17:31:02 <dtantsur> I think RC1 may be even a bit too late for a soft freeze. when we cut stable/pike, it's a hard freeze for it
17:31:05 <rloo_> schedule: https://releases.openstack.org/pike/schedule.html
17:31:18 <rloo_> dtantsur: week before that?
17:31:35 <dtantsur> ah, wait, final releases and the branching is probably on Aug 21
17:31:52 <rloo_> dtantsur: right, final rc is aug 21
17:31:54 <dtantsur> still, I'd start the SFF in August (right, a week before)
17:32:20 <dtantsur> any other thoughts on the SFF topic?
17:32:32 <rloo_> dtantsur: i think i'd start soft freeze for features the week of july 31, not a soft freeze that includes bug fixes
17:32:58 <dtantsur> wait, there may be some misunderstading here. bug fixes are always allowed during freezes, no?
17:33:14 <vdrok> yup, I think so
17:33:17 <dtantsur> I only mentioned them for clarity
17:33:17 <rloo_> dtantsur: yeah, but your definition above included bug fixes and docs
17:33:28 <dtantsur> well, they're in "allowed" category
17:33:46 * dtantsur is a bit confused
17:34:09 <rloo_> dtantsur: so for the discussion today (wrt pruning the priorities) shall we say we want to get whatever landed by week of aug 7, or byjuly 31?
17:34:32 <dtantsur> priorities can land after that date and till the branching, I guess
17:34:51 <rloo_> dtantsur: that's what i'm really intersted in, cuz that gives us time frame wrt toknowing what we can reasonably do by then.
17:34:52 <TheJulia> I thought branching occurs from the last release tag
17:35:06 <dtantsur> TheJulia: right, on the week of Aug 21 IIUC
17:35:27 <dtantsur> so, tl;dr: I suggest we limit landing non-priority features starting with Aug 1st
17:35:58 <TheJulia> I agree
17:36:05 <vdrok> seems fine to me
17:36:07 <rloo_> dtantsur: what about priority features? no limit? (maybe i misunderstand)
17:36:30 <wanyen> dtantsur, iLO team will need more time so ug 7 works better for us
17:36:43 <dtantsur> rloo_: no limit for: 1. priorities (that stay after pruning), 2. small vendor-only (no common layer modification) features, 3. bugs, docs, CI
17:36:44 <wanyen> s/ug/aug
17:36:44 <TheJulia> We have generally become very gun shy about landing major features towards release time
17:36:55 <dtantsur> wanyen: more time for what?
17:37:23 <wanyen> we need time to warp up proliantutils and hpsum patch needs more review
17:37:38 <TheJulia> wanyen: proliantutils has to be released this week.
17:37:41 <dtantsur> I suspect the hpsum patch falls under "small vendor-only features", but I don't remember for sure
17:38:12 <rpioso> dtantsur: What's the freeze date for vendor features, such as a new hardware type driver?
17:38:35 <dtantsur> rpioso: new hardware types for *existing* drivers is IMO a part of the driver composition priority
17:38:49 <dtantsur> so it's not affected by the freeze in my opinion
17:38:57 <dtantsur> others may disagree ^^^
17:39:00 <rpioso> dtantsur: Works for me :)
17:39:02 <TheJulia> I agree
17:39:42 <dtantsur> I guess I'll send a formal proposal to the ML, and we can go from there
17:39:59 <TheJulia> I feel like this is more expectation setting than anything else.
17:40:13 <rloo_> dtantsur: ++ for email.
17:40:14 <dtantsur> should we go over all priorities now? or just pick ones we think are at risk?
17:40:37 <TheJulia> I think prune the at risk ones
17:40:40 <rloo_> dtantsur: i think we hsould be able to go over them all?
17:40:44 <TheJulia> s/prune/re-prioritize/
17:40:53 <rloo_> dtantsur: let's start and see how fast we move...
17:41:04 <dtantsur> ok. so, BFV looks good. keeping?
17:41:13 <rloo_> +1 keep
17:41:19 <dtantsur> rolling upgrades
17:41:30 <dtantsur> how much work is there ahead? it still looks doable to me?
17:41:36 <vdrok> rloo_: the main one left is the online migration command?
17:41:39 <rloo_> +1 keep. we've got most of it in, if we don't get rest in it will be awkward.
17:41:48 <dtantsur> yeah, it's going to be weird
17:41:53 <dtantsur> driver composition
17:41:56 <rloo_> 2 patches left for main rolling upgrades. 1 has two +2 but dying cuz of neutron bug.
17:42:00 <dtantsur> this is only docs and new hardware types
17:42:04 <vdrok> ++ to keep then
17:42:09 <TheJulia> +1 keep
17:42:22 * dtantsur is a bit too fast, I guess
17:42:41 <vdrok> I think both of those can be landed up until hard freeze
17:42:46 <dtantsur> ok with the driver composition remaining bits?
17:42:50 <vdrok> (docs and new hw types)
17:42:51 <rloo_> dtantsur: umm, wait. what happened to ci refactoring?
17:43:03 <rloo_> dtantsur: L125
17:43:05 <dtantsur> rloo_: it's not a priority, it's "do it always" thing
17:43:19 <dtantsur> technical-debt-ish
17:43:27 <dtantsur> being not a feature, it's not affected by any freezes anyway
17:43:34 <rloo_> dtantsur: ok, going to note that.
17:44:02 <dtantsur> OSC default API version change - done for Pike, I guess?
17:44:21 <rloo_> dtantsur: what about reference architecture guide? L175
17:44:25 <dtantsur> hmm, or do we want a warning for "ironic" utility?
17:44:41 <dtantsur> oh, sorry, my screen scrolled for some reason
17:45:06 * rloo_ was wondering what list/order dtantsur was using...
17:45:09 <dtantsur> we're doing some ref-arch-related work downstream, so I can well apply some bits upstream
17:45:23 <rloo_> dtantsur: so you think you'll have something for pike?
17:45:30 <dtantsur> yes, I still believe so
17:45:37 <rloo_> dtantsur: great
17:45:44 <dtantsur> other comments?
17:45:55 <rloo_> hmm, is python 3.5 a pike priority? L178
17:45:59 <TheJulia> it is
17:46:04 <dtantsur> yeah, it's a global priority
17:46:07 <rloo_> deadline for that is next week i think
17:46:13 <rloo_> what does it mean if we don't meet it?
17:46:26 <dtantsur> nothing really.. we'll have to finish it in Queens, I guess
17:46:39 <rloo_> dtantsur: ok, in that case, i suspect this may drop
17:47:06 <dtantsur> I'm not sure we can just drop it from our priorities.. but given how much attention we paid to it, I don't really know
17:47:13 <dtantsur> I'm fine with moving it to Queens
17:47:21 <dtantsur> opinions on that?
17:47:48 <rloo_> dtantsur: honestly, we can leave it there but realistically... we could at least make a note that it has lower priority?
17:47:54 <vdrok> are the two patches the only ones required in ironic?
17:48:13 <dtantsur> vdrok: I don't quite know. I have the same information as you, unfortunately
17:48:19 <vdrok> Nisha_Agarwal: are you around?
17:48:31 <Nisha_Agarwal> vdrok, yes
17:48:35 <rloo_> vdrok: there is that bullet about 'other patches not merging' so i think they are blocked
17:49:02 <vdrok> rloo_: yup, tho I think it's about project-config?
17:49:11 <dtantsur> IIRC "other patches" are about not ironic itself
17:49:15 <rloo_> Nisha_Agarwal: what is needed to complete the python 3.5 work? the status is unclear to me.
17:49:19 <dtantsur> we can only reasonably target a python 3 job for ironic
17:49:22 <dtantsur> for Pike
17:49:25 <vdrok> if there are only two patches in ironic itself, we can still leave it
17:49:29 <rloo_> Nisha_Agarwal: are we done if we get those two ironic patches merged?
17:49:38 <Nisha_Agarwal> vdrok, after we have these two patches merged in ironic then we will know if py35 gate passes or still fail
17:50:17 <vdrok> Nisha_Agarwal: mhm, ok. I'll try to help with the first one with hashing then
17:50:20 <rloo_> Nisha_Agarwal: the first patch you have a -1 for workflow so it isn't even ready to be reviewed
17:50:21 <Nisha_Agarwal> because i never saw any other failures other than these failures yet
17:50:48 <vdrok> I'd vote for leaving it for a week, and see if it progresses
17:50:55 <vdrok> if not, we can drop :)
17:51:03 <rloo_> we've got 10 min left, i say lower the priority until Nisha_Agarwal can provide more confidence to it. we can't even review the first patch yet.
17:51:04 <Nisha_Agarwal> i will update that patch by today/tomorrow
17:51:11 <dtantsur> cool, yeah
17:51:14 <TheJulia> If we can review the virtualbmc patch.... and get that merged, we would be in a better position overall
17:51:16 <Nisha_Agarwal> the second one can go in
17:51:26 <Nisha_Agarwal> yes
17:51:47 <rloo_> ok, we can revisit next week meeting Nisha_Agarwal if you feel like it works.
17:51:53 <dtantsur> yeah, let's move on
17:51:57 <dtantsur> Deploying with Apache and WSGI in CI
17:52:07 <dtantsur> done for ironic, will not be done for inspector
17:52:14 <Nisha_Agarwal> Thanks rloo_ it would. It worked in py35 shell
17:52:19 <Nisha_Agarwal> but when i ran gate it failed
17:52:21 <Nisha_Agarwal> :(
17:52:21 <rloo_> is that a pike community goal?
17:52:28 <vdrok> yup
17:52:31 <rloo_> Nisha_Agarwal: please let us know next week, we have to move on
17:52:36 <Nisha_Agarwal> sure
17:52:42 <dtantsur> rloo_: another community goal, yes. inspector is very far still
17:52:54 <rloo_> dtantsur: what's your thought then wrt inspector?
17:53:04 <dtantsur> rloo_: we do it in Queens
17:53:06 <rloo_> dtantsur: I'm thinking punt to Queens
17:53:08 <rloo_> :)
17:53:28 <dtantsur> ok, we discussed the driver composition..
17:53:43 <dtantsur> OSC default API version change - do we need the warning for the "ironic" tool as well?
17:53:45 <rloo_> we need snmp and drac hw types?
17:53:52 <dtantsur> rloo_: and oneview IIRC
17:54:09 <rloo_> but ok, we need to get driver comp done so can discuss later wrt what we need to do
17:54:10 <dtantsur> maybe Cisco stuff, I don't remember
17:54:33 <rloo_> dtantsur: so OSC default API change. we are done til Queens?
17:54:33 * etingof is snmp guru
17:54:48 <dtantsur> rloo_: depending on whether we want a warning for "ironic" tool. the spec mentions it.
17:54:50 <TheJulia> w/r/t a warning for ironic cli, I thought we agreed to it, but I've not seen a patch for it
17:54:56 * dtantsur assigns etingof to everything
17:55:07 * TheJulia +1's this assignment of everything
17:55:09 <dtantsur> TheJulia: I can make a patch quite quickly
17:55:10 <rloo_> dtantsur: the more warnings the better
17:55:15 <TheJulia> dtantsur: awesome
17:55:18 <rloo_> dtantsur: thx.
17:55:44 <dtantsur> Node tags. I haven't followed it, to be honest. how far are we?
17:56:07 <TheJulia> not in a good place
17:56:22 <TheJulia> looks like we will have to punt, although perhaps more substrate can make it in before the end of the cycle
17:56:30 <dtantsur> we took node tags because the spec and the changes were around for some time
17:56:31 <vdrok> 3 patches away :) one of them is api. yeah, mostly trying to decide with rloo_ the way to deal it wrt rolling upgrade
17:56:34 <dtantsur> I don't think we had big asks for it
17:56:49 <dtantsur> I'd prefer vdrok to spend his time on something with "network" in its name ;)
17:56:58 <vdrok> can be lowered in priority, sure
17:57:00 <rloo_> i would like to get it in cuz it has been there for awhile but i don't think it is a high priority.
17:57:11 <rloo_> definitely lower priority than networking.
17:57:21 <dtantsur> yeah, lowering makes sense to me
17:57:40 <dtantsur> now, sigh. rescue mode
17:57:43 <rloo_> 3 minute warning
17:58:02 <dtantsur> it's important (IMO), but it's not moving (again, IMO)
17:58:06 <rloo_> do we know if 1. rescue mode works end to end; 2. how are reviews going?
17:58:36 * vdrok reviewed it like several months ago :(
17:58:50 <dtantsur> I haven't reviewed them ever
17:59:05 <dtantsur> I will have time and prioritization for them, but in Queens
17:59:07 <TheJulia> Once a long time ago, it just hasn't been on the radar really
17:59:08 <rloo_> so again, i would love to have it, but i don't see much 'i want' for this as a higher priority than other features.
17:59:18 <milan> 1 min left
17:59:19 <rloo_> let's lower the priority
17:59:22 <dtantsur> let's lower
17:59:27 <dtantsur> and I guess let's finish on the next meeting?
17:59:41 <rloo_> i think we can whip through some of the others.
17:59:50 <milan> 10s left
17:59:51 <dtantsur> note in 30 seconds :)
17:59:53 <rloo_> post-deploy vif attach/detach is up to nova
17:59:53 <dtantsur> oh
17:59:58 <dtantsur> thanks everyone!
17:59:59 <rloo_> physical network awareness we want
18:00:02 <dtantsur> #endmeeting