14:00:01 <cdent> #startmeeting placement 14:00:02 <openstack> Meeting started Mon Mar 25 14:00:01 2019 UTC and is due to finish in 60 minutes. The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:05 <openstack> The meeting name has been set to 'placement' 14:00:10 <gibi> o/ 14:00:11 <edleafe> \o 14:00:13 <takashin> o/ 14:00:13 <alex_xu> o/ 14:00:16 <cdent> If you're here for the 1st placement meeting called a placement meeting say hi 14:00:39 <edleafe> I was told there would be punch and pie 14:00:42 <mriedem> o/ 14:00:54 <cdent> that comes at the end of the meeting, if you're well behaved 14:01:03 <tetsuro> o/ 14:01:10 <cdent> efried, jaypipes : you here? 14:01:19 <efried> o/ 14:01:24 <bauzas> \o 14:01:27 <cdent> #topic last meeting 14:01:34 <cdent> #link last minutes http://eavesdrop.openstack.org/meetings/nova_scheduler/2019/nova_scheduler.2019-03-18-14.00.html 14:01:40 <bauzas> goodbye nova-scheduler 14:01:53 <cdent> action items were: 14:02:01 <cdent> cdent nova-to-placement upgrade doc 14:02:01 <cdent> mriedem to document compute driver capability traits per https://review.openstack.org/#/c/538498/ 14:02:02 <cdent> cdent make a ptg etherpad with refactoring goals and constraints on it 14:02:14 <cdent> mine were done. mriedem did you do? 14:02:48 <mriedem> i'm having a hard time parsing the question 14:02:54 <mriedem> oh yes 14:02:57 <cdent> your action 14:03:09 <mriedem> https://review.openstack.org/#/c/644293/ 14:03:17 <jaypipes> cdent: yes 14:03:26 <cdent> #link docs for compute capabilities https://review.openstack.org/#/c/644293/ 14:03:34 <cdent> yay, all dos done 14:03:43 <cdent> #topic specs and review 14:03:49 <cdent> #link latest pupdate http://lists.openstack.org/pipermail/openstack-discuss/2019-March/004149.html 14:04:03 <cdent> re specs, we've started the pieces to do specs in tree 14:04:15 <cdent> #link in tree specs https://review.openstack.org/#/c/645195/ 14:04:32 <cdent> would be good if multiple people could review that, since it is a process change 14:04:39 <efried> cdent: Do we have a fast-approval process for placement specs that were approved in stein? 14:04:49 <efried> I'm thinking of the ones tetsuro had 14:04:51 <jaypipes> mriedem: nice work on that compute caps doc. well written. 14:04:58 <jaypipes> mriedem: and thank you. 14:05:10 <cdent> efried: we can, hadn't really thought of it yet. 14:05:18 <efried> swhy I brought it up :) 14:05:51 <cdent> I was sort of hoping that if/when things get proposed for specs that ought to be fast, we'll just be fast and not require an explicit statement of such 14:06:10 <cdent> but yes, tetsuro's already in flight stuff makes sense to get in asap 14:06:56 <mriedem> that's generally how i handle nova spec re-proposals, 14:06:57 <cdent> Is anyone already working on _new_ placement specs? 14:07:09 <mriedem> but they do sometimes hit a snag, like the sev re-proposal for train 14:07:43 <jaypipes> cdent: does the cpu resource tracking one count? 14:07:52 <jaypipes> more nova.. 14:07:59 <cdent> jaypipes: that one is forever young 14:08:53 <cdent> I intend to start writing up actual contributing docs some time this week: how to make a bug, how to start a feature, when to do a spec, etc. In there I'll include notes on re-proposals. Agan those will be something where it would make sense to have lots of review. 14:08:54 <efried> cdent: Working on in my head, nothing written down yet. 14:09:03 <cdent> ✔ 14:10:05 <cdent> tetsuro: are you hoping to un-conflict your negative member of stuff soon? 14:10:25 <tetsuro> I'm thinking doing that in Train cycle 14:10:37 <cdent> we're open 14:10:43 <cdent> well, nearly 14:10:54 <tetsuro> Ah the brunch was cut, okay 14:10:57 <cdent> you can certainly starting working on, but we might want to wait to the end of the week before merging 14:11:17 <efried> there's still a -1 on there 14:11:25 <efried> but yeah, may as well spruce it up 14:11:37 <cdent> speaking of stable/stein there is at least one, maybe two things that will be in rc2 14:11:41 <gibi> I have the any-traits in a_c spec in my list but I don't know when I will get back to it 14:11:41 <cdent> but otherwise it's pretty ... stable 14:11:59 <gibi> as well as the request group mapping in a_c 14:12:32 <cdent> gibi: do you need/want to free it up for someone else, or? 14:12:47 <gibi> cdent: sure both is free to take if somebody wants to start it 14:12:56 <cdent> k 14:13:27 <cdent> A review which is stuck: 14:13:46 <cdent> #link gabbi-tempest https://review.openstack.org/#/q/topic:cd/gabbi-tempest-job 14:14:12 <jaypipes> cdent: I'll revisit this morning. 14:14:30 <mriedem> i'll ask again, 14:14:35 <cdent> we've had general agreement that this is a good idea (since it has found bugs that other things have not), however, it needs more review. efried reviewed it but was reluctant to +2 because of lack of familiarity with zuul and tempest related stuff, if I remember right? 14:14:52 <efried> yes 14:14:54 <mriedem> what does this cover that we don't have with the tempest-full job 14:14:55 <mriedem> ? 14:15:21 <cdent> mysql-based used of the placmenet api, apparently 14:15:27 <cdent> s/used/use/ 14:15:31 <cdent> that was the bug it found 14:15:36 <efried> IMO it was more about setting the precedent of being able to write jobs in gabbi 14:15:38 <mriedem> devstack is running with mysql by default 14:15:52 <cdent> nova apparently doesn't talk to placement in the same way that those gabbits do 14:16:20 <cdent> but yes, what efried said 14:16:34 <cdent> tempest is huge, painful, noisy, inscrutable and slow 14:17:09 <mriedem> umm 14:17:17 <mriedem> i guess i'll just leave comments on the review 14:18:59 <cdent> please do. It's not a done deal, but several people "like it", so... 14:18:59 <cdent> anything else on reviews, specs, etc? 14:18:59 <cdent> #topic bugs 14:19:00 <cdent> #link Placement bugs (launchpad) https://bugs.launchpad.net/nova/+bugs?field.tag=placement 14:19:00 <cdent> #link Placement storyboard https://storyboard.openstack.org/#!/project_group/placement 14:19:13 <cdent> I'm not aware of anything new or pressing. Anyone got anything? 14:19:59 <cdent> #topic PTG/Forum 14:19:59 <cdent> #link etherpad for ptg https://etherpad.openstack.org/p/placement-ptg-train 14:21:37 <cdent> As I said in a recent email, we do have a room at the PTG but limited time and not everyone can be there. Given that, I'm wondering if people are okay with the idea of starting most of the talking, in email, before the PTG? 14:21:54 <jaypipes> cdent: that would definitely be my preference. 14:21:57 <efried> ++ 14:22:42 <edleafe> sounds good 14:23:07 <cdent> I think many people will have to be with nova, blazar, etc, so there won't be much time left over for placement stuff. I'd rather we spend the time in person resolving the hard problems we couldn't figure out over email, doing a bit of retrospective and long term thinking and flavoring, and simply enjoying each others company 14:23:32 <edleafe> that last part might be a stretch :) 14:23:45 <cdent> feh! 14:23:53 <cdent> I'll orchestrate that email 14:23:54 <bauzas> threads are sometimes rat holing, but if the consensus is done by the PTG or by one last email, good then 14:24:30 <bauzas> and I honestly said "sometimes" because I want to be a gentleman 14:24:47 <cdent> related to that: I'll be travelling the week before the PTG. Some email, so I'll be able to summarize before the ptg. 14:24:49 <mriedem> so...pistols at dawn? 14:25:09 <cdent> ratholing in email is at least better than ratholing in person, which would happen anyway, more, without the email 14:25:22 <mriedem> and is not digestable afterward 14:25:34 <mriedem> when we say, "why didn't we go this way 6 months ago when talking about this?" 14:25:52 <mriedem> "oh idk it was at the ptg but nobody took notes" 14:26:12 <bauzas> that's a good point, we sometimes discuss without notes 14:26:24 <bauzas> but I think we rathole more with emails than F2F 14:26:30 <bauzas> but meh 14:26:33 <cdent> ymmv 14:26:39 <mriedem> are we rat holing about rat holing? 14:26:55 <efried> impressive 14:26:55 <cdent> in this particular case we simply won't have the time to rat hole or not, so it is moot: we have to talk beforehand 14:27:01 <cdent> we are so good at it 14:27:11 <bauzas> heh 14:27:13 <bauzas> anywa 14:27:19 <bauzas> anyway, just my thoughts 14:27:35 <cdent> another absence: i'll be at openinfra uk next monday, if somebody else wants to run the meeting, that's cool, if not, we can just skip it (but see later in the agenda0 14:27:54 <cdent> anything else on ptg forum? 14:28:07 <cdent> #topic opens 14:28:23 <cdent> How wed are people to having a meeting? 14:28:53 <mriedem> as in this weekly meeting? 14:28:58 <cdent> yes 14:29:13 <mriedem> i'd be fine with saying it's in place if needed but cancel ahead of time if there is nothing interesting on the agenda? 14:29:17 <gibi> it feels quick so it works for me 14:29:48 <edleafe> I'm much better with it at (my) 9am than 8am :) 14:29:50 <cdent> mriedem: that's one of the options that was discussed. others were: kill it dead, keep it as is, do office hours in #openstack-placement instead (potentially more often) 14:30:53 <mriedem> are office hours in the same meeting format with action items and such? 14:30:58 <cdent> no 14:31:05 <mriedem> i'm thinking about apac people who aren't able to hang out in irc all the time 14:31:05 <cdent> it's just a designated time to discuss 14:31:19 <mriedem> they might also just go to the ML, but idk 14:31:20 <cdent> "we know people will be there" 14:31:38 <mriedem> if this meeting is just an echo chamber then sure kill it 14:31:57 <cdent> well, to me it feels like an echo chamber with a rathole here and there 14:32:07 <mriedem> pupdate emails will be even more important then 14:32:09 <cdent> but I don't want to impose 14:32:15 * cdent nods 14:32:17 <efried> Since it's got a new purse and shoes, let's keep it for a few weeks and reevaluate. 14:32:22 <mriedem> some of us aren't following along in the placement channel all day either 14:33:19 <bauzas> I just feel it's easier for a new contributor to sit down, lurk at the meeting and shout during opens than starting to write a ML thread 14:33:35 <cdent> how about this: we keep it scheduled until the ptg and then we'll decide there. I will miss two between now and then 14:33:35 <mriedem> deciding not to decide now is fine too, things are a bit hectic right now 14:33:38 <bauzas> for office hours, meh I'm not opiniated 14:33:44 <mriedem> cdent: yeah that seems good 14:33:58 <cdent> bauzas: that's entirely different from my experience, which is why I asked for input, so qed, I guess 14:34:15 <bauzas> cdent: yeah I don't disagree 14:34:26 <bauzas> maybe it's different for anyone 14:34:29 <cdent> any other opens? including newcomers who have been lurking and need to shout now? 14:35:54 <bauzas> looks like it's sold 14:36:18 <cdent> In terms of the aformentioned "a bit hectic now" I agree, but most of that is outside of placement. I reckon everyone needs to get back to that and we'll let placement percolate for a while. Thanks for coming everyone. 14:36:33 <cdent> 1 14:36:35 <cdent> ...2 14:36:39 <cdent> #endmeeting