21:01:04 <mikal> #startmeeting nova
21:01:05 <mriedem> i have lots of grievances!
21:01:05 <openstack> Meeting started Thu Mar 26 21:01:04 2015 UTC and is due to finish in 60 minutes.  The chair is mikal. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:01:08 <openstack> The meeting name has been set to 'nova'
21:01:09 <alaski> o/
21:01:11 <edleafe> \o
21:01:14 <dansmith> o/
21:01:19 <sdague> o/
21:01:20 <melwitt> o/
21:01:24 <lennyb> чи
21:01:25 <mikal> #topic Kilo status
21:01:26 <lennyb> Hi
21:01:31 <mikal> Ok, so we're feature frozen
21:01:34 <mikal> And string froken
21:01:39 <mikal> frozen even
21:01:42 <mikal> And deps frozen
21:01:46 <jogo> o/
21:01:47 <mikal> So, just bug fixes now
21:01:54 <mriedem> i watched frozen on monday while snowed in :)
21:02:06 <mikal> mriedem: you were actually frozen? How excitement.
21:02:13 <cfriesen> mriedem: how many times does that make?
21:02:16 <cfriesen> :)
21:02:17 <mikal> So, anything else to say there? Its just a reminder really.
21:02:28 <bauzas> mriedem: welcome in the parents of daughters club
21:02:39 <claudiub> q: novaclient is also frozen?
21:02:51 <mikal> We don't freeze the client, but we might not release it either
21:03:01 <jogo> well we need to have a stable branch for clients now
21:03:09 <mikal> jogo: did that merge yet?
21:03:19 * mikal hasn't looked since the TC meeting about it
21:03:19 <claudiub> k, thanks. :)
21:03:40 <jogo> mikal: its not about things merging, that is the state we are in today
21:03:49 <mikal> So yeah, there is a proposal for stable branches for clients, but I don't think its merged yet unless it happened without me noticing
21:03:59 <jogo> dhellmann: ^
21:04:00 <mikal> jogo: well, the spec for what to do needs to merge, right?
21:04:18 * mikal is happy for us to have a stable branch for novaclient
21:04:19 <claudiub> well, there is at least one commit that will have to be considered before the client stable release
21:04:26 <mikal> I just want to make sure we follow the agreed process
21:04:31 <claudiub> it's related to the microversions, on novaclient
21:04:45 <sdague> claudiub: we're already pinned on clients, so it will require a req bump, which means probably post release
21:04:46 <claudiub> without it, the microversioning api can't be used
21:05:03 <sdague> claudiub: it can be used by people using the API directly
21:05:04 <jogo> mikal: right, the mechanics of it sure
21:05:38 <claudiub> i see.
21:06:04 <claudiub> although, i don't know how many people use the api directly, without using the novaclient. :)
21:06:10 <sdague> anyway, once the client patch lands, we can see where we stand and if it's worth it
21:06:15 <mikal> Herm, the TC meeting agenda has been bumped to next week, so its not trivial for me to get a URL for the spec
21:06:47 <cfriesen> extended server attributes wants microversion 2.3, so if novaclient doesn't support that...
21:06:51 <mikal> Ahhh, this
21:06:53 <mikal> #link https://review.openstack.org/#/c/155072/
21:06:59 <mikal> That's what jogo is referring to
21:07:37 <mikal> Well, if we think there are reviews we need to do for the client, let's track that somehow and then make a call about releasing once they're merged
21:07:50 <mikal> Perhaps a next section on the priority etherpad?
21:07:59 <mikal> Although they could just go in the v2.1 api section...
21:08:30 <claudiub> makes sense to be in the v2.1 api section, as it is a consumer
21:08:35 <mikal> Works for me
21:08:49 <mikal> Anything else on release status?
21:09:12 <mikal> #topic Kilo priorities
21:09:29 <mikal> So, apart frm those client reviews, anything else we need to be trackig for priorities?
21:09:43 <sdague> https://review.openstack.org/#/c/142495/ would be nice to land
21:09:53 <sdague> annegentle has been trying to get the api docs into our repo for months
21:09:56 <dansmith> has a -1
21:10:02 <dansmith> I've been watching for that to bump today
21:10:08 <dansmith> happy to +W when it does
21:10:10 <mikal> Also, https://review.openstack.org/152569 is on the agenda but I don't know who put it there
21:10:17 <sdague> yes, it does for complaints about the doc content
21:10:28 <sdague> but these are our docs
21:10:40 <sdague> and redoing them should happen after move
21:10:55 <dansmith> sdague: you're saying it should land in its current form?
21:11:03 <dansmith> sdague: annegentle seemed to be working on a change
21:11:11 <melwitt> mikal: I think that was me from two weeks ago
21:11:26 <mikal> melwitt: le huh, fair enough
21:11:32 <melwitt> haha
21:11:53 <sdague> dansmith: honestly, I'd be happy with current land, and fix it after.
21:12:06 <sdague> as I expect this to be a -1 to death patch otherwise
21:12:33 <dansmith> sdague: can you comment to that effect/
21:12:40 <sdague> dansmith: sure
21:12:51 <dansmith> mikal: you okay with that?
21:13:05 <mriedem> fwiw i asked kaufer to review the pagination section since he's been pretty involved there
21:13:09 <mikal> dansmith: I am
21:13:19 <dansmith> mriedem: yeah, it's definitely good that he did
21:13:20 <mikal> dansmith: we can iterate in tree
21:13:30 <dansmith> mriedem: and I was waiting because it looked like good changes
21:13:38 <dansmith> but I'm also happy to not hold up docs
21:13:41 <annegentle> I really do have updates, just need to finalize them
21:13:51 <dansmith> annegentle: want us to just wait then?
21:13:55 <mriedem> i'm asking kaufer in -nova if he will make changes after it's merged
21:13:58 <mriedem> but sounds like we should wait
21:14:23 <dansmith> annegentle: you can ping one of the four of us and we can help make sure it lands on the next round if you want
21:14:28 <sdague> annegentle: ok, your call, it seemed like this keeps getting stalled though
21:15:31 <mikal> Sounds like we wait for annegentle then?
21:15:49 <mikal> Anything else on priorities?
21:16:05 <annegentle> thanks dansmith sdague
21:16:10 <mikal> #topic Gate status
21:16:10 <annegentle> I'll have it done by my eod
21:16:12 <bauzas> RC1 planned for April 9th, rihgt?
21:16:21 <mriedem> gate has been ok i think?
21:16:26 <mikal> bauzas: or before, when the targetted bugs are fixed
21:16:34 <mikal> mriedem: yay!
21:16:39 <mikal> #topic Bugs
21:16:43 <mriedem> overall categorization is down though http://status.openstack.org/elastic-recheck/data/uncategorized.html
21:16:44 <mriedem> badly
21:16:45 <bauzas> mikal: okay, I'll update the prio etherpad for related bugs then
21:16:58 <mikal> So, this flows into the RC1 question
21:17:07 <jogo> mriedem: that is because the number of failures is way down
21:17:07 <mriedem> if you filter ^ by 24 hours though there aren't huge missing things
21:17:10 <mikal> RC1 is blocked until the targetted bugs are fixed, or bumped to RC2
21:17:11 <anteaya> mriedem: did you want do do an elastic recheck categorization day?
21:17:21 <jogo> 112 failures in 10 days is not bad
21:17:22 <mriedem> anteaya: no, there isn't anything worth categorizing right now
21:17:24 <mriedem> yeah
21:17:27 <anteaya> mriedem: k
21:17:37 <sdague> bauzas: that's the *last* day for rc1, it might go earlier
21:17:37 <mriedem> i do have this https://review.openstack.org/#/c/168158/
21:17:50 <bauzas> sdague: ack, thanks for the clarification
21:17:54 <dansmith> I was looking and our -rc1 bug list is pretty short
21:17:54 <mriedem> i'm back to digging in the guts of quotas for this fixed_ips race we've had for a year now
21:18:01 <mikal> Yeah
21:18:01 <dansmith> is everything we really care about on that?
21:18:04 <mikal> #link https://launchpad.net/nova/+milestone/kilo-rc1
21:18:21 <dansmith> I expected there to be some bugs that weren't in-progress on there, but there really isn't much
21:18:25 <bauzas> I have a few bugs I'm concerned that are targeted for rc1 but that's on-going
21:18:28 <mikal> So I think last release what we did about now is have a rat around looking for bugs we thought should be RC
21:18:47 <mikal> I don't think we should be randomly promoting things
21:18:56 <mikal> But if people know about something important, they should make sure its on that list
21:19:10 <dims> like quotas!
21:19:11 <bauzas> +1 for a bug triage
21:19:11 <sdague> +1
21:19:12 <cfriesen> any chance of getting someone to look at https://review.openstack.org/#/c/164762 ?  should be fairly uncontroversial.  It hasn't gotten much review traffic.
21:19:23 <dansmith> yeah, I just have this feeling like maybe there isn't much on there, and maybe it's not getting focus
21:19:35 <sdague> dims: quotas is bigger than just some bug fixes
21:19:50 <dims> sdague: true
21:20:10 <mikal> Anyway, we don't have to find all the rc bugs right now in this meeting
21:20:10 <jogo> sdague: there are a few self contained bugs though
21:20:15 <mikal> But we do need to make sure we think about it
21:20:35 <mriedem> the cells job is non-voting but -1 for a few weeks,
21:20:36 <jogo> it may be worth taking a look at the logs to see if they look sane too
21:20:46 <mriedem> are all of the patches up to make that green? if not, we should have rc1 bugs for those maybe?
21:20:55 <bauzas> mriedem: we're actively working on fixing the issues with the cells job
21:20:55 <mriedem> jogo: right - same idea ^
21:21:02 <cfriesen> sorry folks, that was the wrong link...that's the trickier one.   I meant to ask about https://review.openstack.org/#/c/162746/
21:21:11 <bauzas> mriedem: the main ones are already targeted RC1
21:21:25 <mriedem> bauzas: ok, do any of those have a green cells job at the HEAD?
21:21:53 <sdague> mriedem: it's 13 fails
21:21:55 <melwitt> mriedem: no because a couple of project-config changes are needed, which I have up depending on bauzas fixes
21:21:58 <alaski> mriedem: https://review.openstack.org/#/c/166396/2 should end us with a green job
21:22:01 <bauzas> mriedem: we need to sync up with melwitt and alaski
21:22:07 <sdague> 10 go away when bauzas updates his patches
21:22:35 <bauzas> sdague: yeah, I expect it to be uploaded by tomorrow morning my time
21:22:38 <mriedem> ah cool, more skips
21:22:48 <melwitt> mriedem: one more skip, many removals :P
21:22:53 <sdague> alaski: we shouldn't exclude the hypervisor tests
21:23:04 <sdague> because that's a reall stacktrace in cells
21:23:06 <bauzas> yeah thanks to melwitt, we'll skip those bad
21:23:08 <mriedem> btw i'm not bashing, just saying it'd be nice to see that green before release
21:23:27 <bauzas> mriedem: yeah that's our target, agreed
21:23:32 <alaski> sdague: I don't think those are being skipped, but I'll double check
21:23:33 <dims> +1 to green
21:23:36 <melwitt> sdague: they won't be, the patch series that's up will actually fix those hypervisor things without excluding them
21:24:03 <mriedem> is that this? https://review.openstack.org/#/c/167564/
21:24:23 <sdague> alaski: oh, maybe it was an old rev
21:24:24 <bauzas> mriedem: not this one exactly, but the series yes
21:24:28 <melwitt> mriedem: yes. and then I have https://review.openstack.org/#/c/166396/ Depends-On that patch
21:24:35 <alaski> sdague: gotcha
21:24:42 <mriedem> yeah that's what i saw
21:24:43 <sdague> yeh, you linked /2 :)
21:24:50 <sdague> and /3 didn't have it
21:24:52 <melwitt> oh, oops
21:25:03 <sdague> ok, +A on that
21:25:12 <alaski> oh, woops
21:25:14 <bauzas> melwitt: you will probably have to change the Depends-On tag, but let's sync us offline
21:25:25 <melwitt> bauzas: okay
21:25:34 <mikal> Anything else on bugs?
21:25:43 <mriedem> there was talk of a bug day
21:25:48 <mriedem> but after rc :)
21:25:55 <mikal> The Shanghai thing?
21:25:56 <sdague> got to drop early, our room is getting abandoned here. Catch you later folks
21:25:56 <mriedem> yeah
21:26:02 <mikal> Yeah, the timing there is going to be a problem
21:26:18 <mriedem> huawei is going to let us pair program in china for 3 bug days
21:26:18 <mikal> So, let's let that sort itself out on the thread
21:26:25 <anteaya> sdague: :( I was hoping to figure out what you wanted to do about the migration
21:26:45 <mikal> #topic Stuck reviews
21:26:57 <mikal> Noting that we're constrained in what we are willing to merge at the moment...
21:27:00 <mikal> Any more stuck reviews?
21:27:09 <mikal> i.e. reviews core will never reach concensus on in gerrit?
21:27:23 <mriedem> so cfriesen was bringing up the server group thing but that's just lack of review it looks like
21:27:32 <mikal> Yep
21:27:34 <cfriesen> yes
21:27:34 <mriedem> i'm not super familiar with that code, looks racy
21:27:37 <mriedem> he's trying to fix it
21:27:41 <mriedem> i don't know how well it's tested
21:27:42 <tonyb> mikal: I've seen one but it's an L thing now so ....
21:27:49 <mikal> tonyb: ok
21:28:12 <mikal> Ok, cool
21:28:17 <mikal> #topic Open Discussion
21:28:23 <mikal> Let me start by saying I am off work today
21:28:26 <mikal> I'm going camping!
21:28:30 <anteaya> mikal: yay
21:28:34 <mikal> But I am sure y'all will get more done if I am away
21:28:40 <melwitt> this review is semi-stuck i.e. waiting for input from danpb https://review.openstack.org/#/c/151953/
21:28:46 <anteaya> mikal: so yeah as I said above what are we doing about the migration?
21:28:48 <mriedem> ...and he was never heard from again...
21:28:49 <cfriesen> sounds fun...where to?   process question for L...for the purposes of whether a blueprint needs a spec, do "meaningful" flavor extra-specs values count as an API change?
21:29:00 <anteaya> mikal: I could use a bit of guidance
21:29:15 <mikal> melwitt: that review has a lot of +2's
21:29:41 <mikal> anteaya: yeah, fair enough.
21:29:50 <mikal> anteaya: as discussed, I think we need a summit session on this yet again
21:29:54 <tonyb> mikal: but nothing from DanB and it's libvirt ....
21:30:02 <anteaya> well that's fine but in the meantime?
21:30:06 <mikal> anteaya: I don't know how we resolve that some operators just aren't excited by neutron
21:30:11 <anteaya> what do I tell oleg just cancel the meetings?
21:30:26 <mikal> anteaya: I think at this point in the release cycle we are unlikely to make much progress
21:30:36 <mikal> anteaya: so putting the meetings on pause for a couple of weeks is a good idea
21:30:36 <anteaya> mikal: how operators felt about neutron was never the problem I was trying to solve
21:30:46 <anteaya> but it seems to have trumped the problem I was
21:30:50 <mikal> anteaya: true
21:30:56 <mikal> anteaya: well, nova wants to delete that code
21:31:00 <mriedem> well, if no one wants to migrate...
21:31:02 <mikal> anteaya: that's why we care about migrating people
21:31:04 <anteaya> that was what I was trying to support
21:31:08 <bauzas> mikal: IIRC there was a question on the last meeting about the planning of the sessions, any chance you could open an etherpad for starting to gather ideas ?
21:31:15 <mikal> anteaya: if people refuse to migrate, we don't get to our code delete end goal
21:31:19 <anteaya> getting to the point of having that code out of nova
21:31:22 <dims> when do we start planning for dev summit sessions?
21:31:25 <anteaya> mikal: agreed
21:31:27 <mikal> bauzas: oh sure, I can to that today
21:31:43 <melwitt> mikal: I know -- it's not controversial but it is stuck waiting for a specific person. just wanted to mention it
21:31:43 <mikal> dims:  later today apparently
21:31:45 <bauzas> mikal: awesome thanks
21:32:00 <mriedem> can nova-network be moved out of tree into some other project/service, like ec2?
21:32:04 <tonyb> dims, bauzas proibably after the PTL election
21:32:11 <anteaya> mikal: well I can post to the mailing list and cancel meetings but I would like to offer something by way of a new direction
21:32:14 <dansmith> mriedem: heh
21:32:29 <mriedem> i'm playing devil's advocate here
21:32:32 <anteaya> mikal: as folks will just batter you with questions if it isn't clear
21:32:38 <mikal> anteaya: if I recall the mail thread correctly we were waiting for a summary of what people said at the ops meetup
21:32:42 <dansmith> mriedem: it's been suggested
21:32:46 <mikal> mriedem: I actually like that idea
21:32:50 <mriedem> well if we don't want it in tree,
21:32:51 <anteaya> mikal: it is a private email thread
21:32:52 <mriedem> but people want it
21:32:54 <mriedem> move it out
21:33:05 <mikal> mriedem: that's one of the thigns for that summit session I think
21:33:21 <dansmith> mriedem: it takes some effort to do that if we want it to work, even for the first day
21:33:22 <anteaya> mikal: and if you are refering to the public one, who is doing a summary and what will it tell us?
21:33:30 <dansmith> mriedem: so we could for sure, but it does take some effort
21:33:40 <anteaya> mikal: I'm fine waiting if I know what I'm waiting for
21:33:40 <mriedem> yeah i don't expect it to be trivial
21:34:00 <dansmith> the other problem is that I think we do piss off some actual people by letting n-net fall into disrepair
21:34:01 <mriedem> dansmith: so like 2 days heads down from you then? :P
21:34:14 <dansmith> heh
21:34:26 <dansmith> anyway, not something we'll settle here
21:34:30 <anteaya> mriedem: I'm actually infavour of your suggestion
21:34:34 <mikal> anteaya: so we're waiting for two things, right?
21:34:44 <anteaya> it seems like the only way forward for some operators
21:34:51 <mikal> anteaya: a migration implementation which doesn't cause concern for nova core
21:34:53 <dansmith> anteaya: it really doesn't solve anything
21:35:00 <mikal> anteaya: and a decision about if we just spin nova-net out into its own thing
21:35:09 <anteaya> mikal: I can live with that
21:35:21 <anteaya> mikal: how do we get those things?
21:35:27 <mikal> anteaya: with the first one probably changing if the second one happens
21:35:33 <anteaya> yes
21:36:23 <mikal> anteaya: I don't have a solution apart from talking through the nova-net split out options at the summit. I need core focussed on the kilo release at the moment, I have a hard deadline for that.
21:36:33 <anteaya> oh of course
21:36:35 <dansmith> +1 for talking about it later
21:36:44 <anteaya> I'm just trying to figure out what to do now
21:36:52 <anteaya> calling a halt doesn't feel right
21:36:57 <mikal> anteaya: I think we have to pause
21:37:02 <mikal> anteaya: I can't think of any other workable option
21:37:11 <mikal> anteaya: the people we need to talk it through with are too busy on the release
21:37:15 <anteaya> mikal: are you able to help me compose a post to the mailing list?
21:37:23 <mikal> Sure
21:37:24 <anteaya> yup, I understand that
21:37:26 <anteaya> thank you
21:37:29 <mikal> We can do that in an etherpad after this
21:37:32 <anteaya> sure
21:38:11 <mikal> Anything else for open discussion?
21:38:13 <lennyb> Hi all, if we are on open discussion, is it a good timing to ask permission to add non-voting Mellanox Nova CI...?
21:38:16 <mikal> Or should I send you all home?
21:38:27 <bauzas> s/home/bed
21:38:29 <mikal> lennyb: jogo had questions for you in the email thread earlier today
21:38:34 <mikal> lennyb: have you seen those yet?
21:38:41 <anteaya> lennyb: you need to bring your link with you
21:38:44 <anteaya> #link http://lists.openstack.org/pipermail/openstack-dev/2015-March/058779.html
21:38:53 <cfriesen> for the L release, for the purposes of whether a blueprint needs a spec, do "meaningful" flavor extra-specs values count as an API change?
21:39:15 <cfriesen> just wondering if I should start writing a spec for a proposal I'm working on
21:39:32 <mikal> What does the word meaningful mean there?
21:39:41 <lennyb> yes, I didnt send him an answer yet.
21:40:00 <cfriesen> stuff that impacts guest topology....I'm looking at allowing explicit pinning of guest numa node to host numa node
21:40:06 <mikal> cfriesen: but, if its part of our public API, it probably needs a spec
21:40:15 <cfriesen> so "hw:numa_node.X=Y" in flavor extra specs
21:40:39 <lennyb> anteaya: thanks.
21:40:43 <mikal> cfriesen: I think that deserves a short spec, so that we don't end up with a syntax we regret later
21:40:56 <cfriesen> okay
21:41:22 <anteaya> lennyb: so whether you answer the question in this meeting, in the nova channel or on the email, the permission you seek is waiting on your response to jogo's questions
21:42:21 <mikal> lennyb: yeah, we need that email answered first please
21:42:27 <mikal> I think that we're pretty much done here?
21:42:45 <mikal> .
21:42:58 <mikal> .
21:43:01 <mriedem> bye!
21:43:06 <mikal> .
21:43:07 <neiljerram> Happy camping
21:43:08 <lennyb> anteaya, mikal: thanks, I will answer the email.
21:43:09 <mriedem> don't get eaten by a wallaby while camping
21:43:13 <mikal> #endmeeting